Clinical Psychology Review 31 (2011) 122–137 Contents lists available at ScienceDirect Clinical Psychology Review Differences between trait fear and trait anxiety: Implications for psychopathology Patrick Sylvers ⁎, Scott O. Lilienfeld, Jamie L. LaPrairie Emory University, United States a r t i c l e i n f o Article history: Received 19 January 2010 Received in revised form 4 August 2010 Accepted 6 August 2010 Keywords: Fear Anxiety Emotion Phobia Panic a b s t r a c t Fear and anxiety are poorly delineated in much of the clinical and research literatures. Although some theorists and researchers have posited explanations for how trait fear and trait anxiety differ, many others conceptualize the constructs as largely or entirely interchangeable. The primary goals of this review are to examine clinical conceptualizations and neurobiological studies of fear and anxiety, examine the animal and human literatures on the correlates of fear and anxiety, provide clearer definitions of these two constructs, and discuss their implications for psychopathology. A secondary goal is to evaluate content of self-report measures of trait fear and anxiety, and meta-analyze the relations between self-reported trait fear and anxiety. We found that existing measures share significant content overlap across constructs. Despite this overlap, our meta-analysis revealed only a moderate (r = 0.32) relationship between measures of trait fear and anxiety, with an even lower relationship (r = 0.14) when we examined trait fear measures operationalized in terms of harm avoidance. These findings suggest that fear and anxiety are largely distinct emotions, and that psychological disorders of trait fear and trait anxiety warrant classification in separate higher-order categories. Moreover, they suggest that future research should focus on deriving more content valid measures of trait fear and trait anxiety from the neurobiological and diagnostic literatures. © 2010 Elsevier Ltd. All rights reserved. Contents 1. Fear and anxiety: Conceptual and definitional confusion . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Conceptual distinctions between fear and anxiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Clinical conceptualizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Neuroscientific conceptualizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Neuroscience research on fear and anxiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Animal studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1. State fear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2. State anxiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3. Trait fear and trait anxiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.4. Implications of non-human animal neuroscience studies . . . . . . . . . . . . . . . . . . . 3.2. Human studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. State fear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. State anxiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3. Trait fear and trait anxiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4. Implications of human neuroscience studies . . . . . . . . . . . . . . . . . . . . . . . . 4. Operationalizations of fear and anxiety: Consensus and disagreement . . . . . . . . . . . . . . . . . . . . 4.1. Defining fear. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Defining anxiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Diagnostic categories of pathological fear and anxiety: Descriptive psychopathology and behavior-genetic findings . 6. Evaluation of the content of self-report fear and anxiety measures . . . . . . . . . . . . . . . . . . . . . . . 6.1. Self-report measures of trait fear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.1. Phobic fear measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 124 124 125 125 125 125 126 126 127 127 127 127 128 128 128 128 128 129 129 129 129 ⁎ Corresponding author. Emory University, Department of Psychology, 36 Eagle Row, Suite 280, Atlanta, GA 30322, United States. Tel.: + 1 404 727 0561; fax: + 1 404 727 0372. E-mail address: psylver@emory.edu (P. Sylvers). 0272-7358/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.cpr.2010.08.004 P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 6.1.2. Harm avoidance measures. . . . . . . . . . . . 6.1.3. Self-report measures of trait anxiety . . . . . . . 7. Meta-analysis of self-report fear and anxiety measures . . . . . . 7.1. Hypothesis 1: Primary meta-analysis . . . . . . . . . . . 7.2. Hypothesis 2: Self-report measure as moderator . . . . . 7.3. Exploratory analyses: Age, gender, and psychiatric status as 8. Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1. Literature search and selection criteria . . . . . . . . . . 8.1.1. Inclusion criteria . . . . . . . . . . . . . . . . 8.1.2. Search results. . . . . . . . . . . . . . . . . . 8.2. Data analytic strategy . . . . . . . . . . . . . . . . . . 8.2.1. Computing effect sizes . . . . . . . . . . . . . 8.2.2. Effect-size aggregation . . . . . . . . . . . . . 8.2.3. Categorical moderator analyses . . . . . . . . . 9. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1. Hypothesis 1: Primary meta-analysis . . . . . . . . . . . 9.2. Hypothesis 2: Self-report measure as a moderator. . . . . 9.2.1. Fear measure as a moderator . . . . . . . . . . 9.2.2. Anxiety measure as a moderator . . . . . . . . . 9.3. Exploratory analyses: Age, gender, and psychiatric status as 9.3.1. Age as a moderator . . . . . . . . . . . . . . . 9.3.2. Gender as a moderator . . . . . . . . . . . . . 9.3.3. Psychiatric status as a moderator . . . . . . . . 9.4. Summary of our meta-analytic findings . . . . . . . . . . 10. Discussion and implications. . . . . . . . . . . . . . . . . . . 10.1. Implications for diagnostic classification systems . . . . . 10.2. Clinical implications . . . . . . . . . . . . . . . . . . . 10.3. Implications for neuroscience . . . . . . . . . . . . . . 10.4. General summary and future directions . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . moderators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . moderators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fear and anxiety are primitive emotions that function to preserve life. In healthy individuals, fear and anxiety can facilitate action in an effort to maintain safety and well-being (Lang, Davis and Öhman, 2000). Nevertheless, when these often adaptive states become characteristic of an individual, or trait-like, the resulting phenomena can turn physically (e.g., cardiovascular disease) and psychologically (e.g., agoraphobia) destructive. Currently, psychopathology marked by high levels of trait-like fear and anxiety reside largely under the umbrella category of anxiety disorders in the diagnostic nomenclature (American Psychiatric Association, 2000). Kessler, Chiu, Demler, and Walters (2005) found that the 12-month prevalence estimate for these common disorders was 18.1%. Moreover, these disorders account for approximately 32% of all mental health expenditures in the United States (DuPont et al., 1996). 1. Fear and anxiety: Conceptual and definitional confusion Despite the severity and pervasiveness of anxiety disorders, the literature is only beginning to clarify the often poorly delineated relationship between trait fear and anxiety. Moreover, the definitional boundaries of trait fear and anxiety remain controversial. Although some theorists and researchers (e.g., Barlow, 2002) have posited explanations for how trait fear and anxiety differ, many others conceptualize the constructs as largely or entirely interchangeable (e.g., Beck & Emery, 2005). For example, Beck and Emery (2005, p. 9) wrote that, “Fear involves the intellectual appraisal of a threatening stimulus; anxiety involves the emotional response to that appraisal.” Izard and Ackerman (2000, p. 260) wrote, “[Fear] is the key emotion in the anxiety pattern.” Some prominent authors have even argued explicitly that fear and anxiety are essentially isomorphic; for example, Wolpe (1987) noted that he chose to use “‘fear’ and . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 130 130 130 130 130 131 131 132 132 132 132 132 132 132 132 132 132 133 133 133 133 133 133 133 134 134 134 134 134 134 ‘anxiety’ synonymously in this paper because they are psychophysiologically indistinguishable” (p. 135). Instead, and consistent with a growing body of research across multiple domains, we contend that there is now persuasive evidence that fear and anxiety are different emotions. We argue that their conflation in the literature has led to ambiguous and at times misleading findings. For example, there is a robust literature on the underpinnings of pathological fear in children (e.g., Hayden et al., 2007). Nevertheless, many of these studies used tasks designed to measure worry, hypervigilance, and rejection sensitivity as proxy measures of fear. As we will discuss later, these tasks appear primarily to assess anxiety, not fear. The confusion is most prominent when investigators use harm avoidance (HA) measures, which can refer to either trait anxiety (e.g., Cloninger, Przybeck, & Svrakic, 1991) or trait fear (e.g., Tellegen, 1982). In research on psychopathy, a number of researchers have attempted to test etiological models positing a lack of fear as the core cause of this condition (e.g., Lykken, 1995) using either measures of trait anxiety or measures that are highly contaminated with trait anxiety. Consequently, some studies that appear to bear on the low fear hypothesis of psychopathy may be largely or entirely irrelevant to this hypothesis (Poythress et al., 2008). Taken together, the conceptual confusion as to fear and anxiety may negatively impact the implications of existing research. The conflation of fear and anxiety, however, is not restricted to the clinical literature. For example, Crestani et al. (1999), in a study investigating the neural pathophysiology of anxiety in rodents, wrote that, “Anxiety states in humans are characterized by harm avoidance behavior and a bias for threat cues” (p. 833). As discussed later, the preponderance of evidence suggests that avoidance behaviors (that is, fight, flight, or freeze responses) are characteristic of fear, not anxiety. As another 124 P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 example, when describing fMRI studies, Davidson (2002) wrote that “One common strategy for evoking anxiety among anxious patients in the laboratory is to present them with specific types of stimuli that are known to provoke their anxiety (e.g., pictures of spiders for spider phobics)” (p. 69). Yet spider phobics are presumably pathologically fearful of spiders and motivated to avoid them; therefore, pictures of spiders may instead provoke flight or freeze responses characteristic of fear, not anxiety. From these examples, it is evident that the relations between trait fear and trait anxiety and their definitional boundaries require clarification. Well-defined trait emotions, however, require clear operationalizations at the state level. Although the phenomena of state fear and anxiety ostensibly share many functional and structural similarities, they may encompass distinct emotional states. In the animal literature, there is a rich discourse examining the biological underpinnings of state fear and anxiety. This literature is clinically relevant given that many of the same structures (e.g., amygdala) that mediate fear and anxiety in animals also mediate these emotions in humans (Blair, 2003). Moreover, as methodologists and philosophers of science have long noted, conclusions are strengthened to the extent to which they derive from maximally diverse sources of information (Shadish, Cook & Campbell, 2001). Several researchers have also recently examined the biological substrates of state fear and anxiety in humans. The findings from these literatures, to be discussed later, suggest that although there is some overlap, there are sharp differences in the patterns of brain activation. In healthy adults, complementary literatures have also noted differential external correlates of self-reported fear versus anxiety, such as pain perception (Rhudy & Meagher, 2000), whereby self-reported fear is inversely and self-reported anxiety positively related to pain perception. This finding is consistent with conceptualizations of fear as an emotion that motivates avoidance behaviors, wherein the organism experiences less physical pain during the fight, flight, or freeze response. This finding is also consistent with the conceptualization of anxiety as an emotion that motivates hypervigilance, wherein the organism exhibits heightened sensitivity to external stimuli during approach. Similarly, several studies using cognitive laboratory tasks have found evidence differentiating fear from anxiety (see Öhman, 2008 for a review). Despite an abundance of biological and cognitive research separating fear from anxiety, definitions of the subjective human experiences of trait fear and anxiety remain contentious (Barlow, 2002). As Gray and McNaughton (2000) noted, however, the only direct means of understanding and defining the subjective experience of fear or anxiety is by self-report. Although preliminary efforts have been made to investigate the discriminant validity of selfreported trait fear and anxiety measures (e.g., Perkins, Kemp, & Corr, 2007), there are no comprehensive reviews of the relations between self-report measures of fear and anxiety, as well as potential moderators of this still murky relationship. The primary goal of this article is to provide a broad overview of research on the clinical and biological correlates of fear and anxiety from both non-human animal and human literatures. In the case of the latter, we examine research on both “normal-range” personality and psychopathological disorders, especially anxiety disorders. We also examine implications of differences between fear and anxiety for the classification of psychopathological disorders and neuroscience research. A subsidiary goal of this paper is to evaluate the content of self-report measures of trait fear and anxiety and metaanalyze their relationships. As discussed later, the meta-analysis is not intended to provide a definitive evaluation of the relationship between self-reported trait fear and anxiety, as we later contend that many of the extant measures are fundamentally flawed. However, we conducted the meta-analysis not only to investigate the associations among these measures but to clarify potential sources of ambiguity inherent in these measures. To set the stage for our narrative review and meta-analysis, we first summarize the theoretical and biological literatures on fear and anxiety, highlighting the most pertinent writings and studies. 2. Conceptual distinctions between fear and anxiety State emotions refer to affective adaptations to specific situations, whereas trait emotions refer to affective characteristics of a person across time and situation (Morrissette, Tull, Gulliver, Kamholz, & Zimering, 2007). Clinical conceptualizations of fear and anxiety are concerned primarily with trait emotions, whereas neuroscientific conceptualizations are concerned primarily with state. Although most neuroscience studies focus on emotion states, their findings are readily extrapolated to form trait operationalizations. 2.1. Clinical conceptualizations In the clinical literature, there are several, in some instances competing, conceptualizations of fear and anxiety. Beck and Emery (2005), for example, posited that fear is a cognitive response to threat, whereas anxiety is an emotional response to fear. In other words, anxiety is the emotional byproduct of fearful cognitions. According to Beck and Emery, trait fear results from pervasive and persistent interpretations of stimuli as threatening. Trait anxiety, in contrast, encompasses the emotional responses to pervasive fearful cognitions. Accordingly, a trait fearful person would presumably be trait anxious. Epstein (1972) suggested that fear is an emotional response that results from the interpretation of specific environmental cues as threatening and manifests itself in avoidance and escape behaviors. He posited that anxiety, in contrast, is a product of one of three different pathways. In the first, avoidance of a feared stimulus is disrupted, a phenomenon termed unresolved fear. Several theorists agree that the inability to avoid fearful stimuli leads to anxiety (e.g., Öhman, 2008). The second, according to Epstein, occurs when individuals overestimate the potential for threat in ambiguous situations. Because the threat is non-specific and future oriented, there are no clear avoidance options, leading to apprehension and indecision. Many contemporary clinical theorists (e.g., Barlow, 2002) define anxiety similarly. Some learning theorists (Zinbarg & Mineka, 2007) support this pathway by suggesting that pathological anxiety results from implicitly associating a benign feature of a dangerous experience with the danger itself. The third type occurs when an individual's environment and expectancies do not match. In terms of behavior, Epstein proposed that anxiety elicits hypervigilance in the face of uncertainty. By Epstein's definitions, trait fear results from an individual being hypersensitive to several distinct environmental threat cues and avoiding situations involving those threats. Trait anxiety, in contrast, results from an inability to avoid a prolonged feared situation, overestimating the potential for threat across situations, or experiencing chronic mismatches between the environment and expectations. According to this model, trait fear and anxiety are largely unrelated, as the specificity of perceived danger cues and their proposed pathways differ. In a review of the cognitive and psychophysiological literature on fear and anxiety, Öhman (2008) posited that fear occurs when individuals are actively coping with a perceived threat, whereas anxiety results from a threatening situation without an effective means of coping. Similar to Epstein's conceptualization of anxiety as unresolved fear, Öhman suggested that fear and anxiety share similar underlying processes and are differentiated based on perceived avoidance options. He also implied that trait fearful individuals who are concerned primarily with physical threat react more strongly than healthy individuals to physically threatening stimuli. Nevertheless, these individuals do not exhibit elevated trait anxiety as they actively avoid perceived threats. Conversely, individuals whose trait fear is concerned primarily with social threat also show elevated levels of P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 trait anxiety. Therefore, the relationship between trait anxiety and fear, according to Öhman's conceptualization, differs according to whether the fear is primarily physical or social (see also Lykken, Tellegen, & Katzenmeyer, 1973). Tellegen (1982) conceptualized trait anxiety as aversive arousal in uncertain situations where avoidance does not seem possible. In contrast, he conceptualized trait fear as hypersensitivity to danger cues leading to avoidance behavior. Using these definitions, Tellegen found that trait fear (as assessed by a “Harm avoidance” scale) and trait anxiety (as assessed by a “Stress Reaction” scale) are separable and nearly orthogonal constructs in the development of the Multidimensional Personality Questionnaire (MPQ; Tellegen, 1982). Tellegen (1985) also found that, when using these definitions, trait fear loads on a higher-order Constraint factor, whereas trait anxiety loads on a higher-order Negative Emotionality factor (see also Watson, Clark, & Harkness, 1994). Individuals who score highly on the Constraint factor “convey caution, planfulness, a tendency to avoid danger, conventionality, and adherence to traditional values” (Tellegen, 1982, p. 36). Individuals who scored high on the Negative Emotionality factor, in contrast, “describe themselves as often stressed and harassed, prone to respond with strong negative emotions to everyday vicissitudes, and as enmeshed in adversarial relationships” (Tellegen, 1982, p. 35). The finding that measures of trait fear and anxiety load on two different higher-order factors of the “Big Three” personality dimensions suggests that fear and anxiety may relate to, and perhaps stem from, different etiological processes (Tellegen & Waller, 2007). 2.2. Neuroscientific conceptualizations Compared with the clinical literature, the constructs of fear and anxiety are delineated more clearly in the neuroscience literature. Fear is commonly defined as an aversive reaction elicited by the perception of a specific threat stimulus, whether conditioned or not (e.g., Cooper & Guynn, 2006). Mammals display three behavioral responses to fear: fight, flight, and freezing in place (Gray & McNaughton, 2000). In most studies of fear reactivity, the threat to the animal or individual is physical (e.g., Berg & Davis, 1985). However, White and Depue (1999) suggested that specificity and degree of threat, not the type of threat, elicit fear. In other words, social situations may also induce fear in humans, as social harm occurs more often and can be equally as devastating as physical harm. Moreover, as Öhman (1986) noted, social threat can quickly escalate to physical harm. Anxiety, in contrast, is commonly defined in the neuroscience literature as prolonged hypervigilance in anticipation of, or response to, a diffuse threat where danger is not clearly imminent (e.g., Macleod & Rutherford, 1992). As Blanchard and Blanchard (1990) noted, however, the perceived imminence of danger in both fear and anxiety is in part a cognitive construction. Therefore, individuals' perception of the specificity and imminence of danger may shape their emotional experience. In other words, perception of danger is subjective and can be influenced substantially by individual differences in personality. Another theoretical distinction is the duration of arousal, whereby the fear response tends to be short lived and the anxiety response long-lived (e.g., Davis, 1998). Trait fear, by these definitions, results from an organism chronically engaging in fight, flight, or freezing behaviors due to perceiving specific environmental cues as threatening. Trait anxiety, in contrast, results from an organism being in a chronic state of hypervigilance due to the anticipation of a generalized threat. By these definitions, trait fear and anxiety are not necessarily related (see also Barlow, 2002). Additionally, McNaughton and Corr (2004), based largely on the works of Jeffrey Gray (1982) and Blanchard and Blanchard (1990), posited a two-dimensional behavioral defense system, whereby fear 125 and anxiety are distinct defensive responses. McNaughton and Corr posited that the certainty of threat is necessary but not sufficient for differentiating fear from anxiety. To McNaughton and Corr, the primary difference between fear and anxiety lies in the directional motivation of the behavior. They described fear as a defensive reaction during which fight/flight/freeze responses facilitate escape from threat. In contrast, they described anxiety as hypervigilance while approaching a potential threat, in the form of physical danger, failure, or loss of reward. Their theory holds that the anxiety system continually compares the current environment with goals and expectations. When conflicts arise between them (approach/approach, approach/avoidance, or avoidance/avoidance), the anxiety system interrupts behavior and places the individual in a heightened state of hypervigilance. In contrast with other neuroscientific conceptualizations, their operationalization of anxiety necessitates conflict arising from competition among available goals. Taken together, the clinical and neuroscience literatures suggest several experiential characteristics of fear and anxiety (see Table 1). Although fear and anxiety appear to share some defining characteristics (e.g., negative valence), they also exhibit several discriminating characteristics (e.g., defensive direction). 3. Neuroscience research on fear and anxiety Despite the relative consistency of neuroscientific conceptualizations, the substrates of fear and anxiety remain the subject of many research endeavors. Fear and anxiety are primitive states, which provide adaptive survival responses to threat (Porges, 1995). The brain structures implicated in fear and anxiety are similar across species, which allows investigators to extrapolate provisionally biological findings from non-human mammals to humans. Although there are species-specific fearful and anxious behaviors, translational research is consistently supported by complementary human studies. The majority of animal research in this area has been conducted using rodents (e.g., Nader & LeDoux, 1999), although some studies used non-human primates (e.g., Winslow, Noble, & Davis, 2007). 3.1. Animal studies 3.1.1. State fear As Gray and McNaughton (2000) pointed out, the only feasible way to infer causal pathways in the neural circuitry related to fear and anxiety is through non-human animal experimentation. The human brain displays comparable connectivity to other mammals with analogous functioning in many areas implicated in fear and anxiety (Aggleton, 1992). Over several decades, neuroscientists have methodically pinpointed the structural and functional substrates necessary to induce fear- and anxiety-associated behaviors in animals, primarily through brain stimulation, lesion, and pharmacological blockade studies in rats. Within these studies, the two most widely investigated behavioral responses include the freezing response and fear-potentiated startle reflex (Brown, Kalish, & Farber, 1951) acquired through classical (Pavlovian) conditioning (Fendt & Fanselow, 1999). A limitation of the freezing response is that it is limited by a zero baseline (Lang et al., 2000), meaning that it only occurs during a fear state. Therefore, it is unclear in pharmacological and lesion studies whether the effect of a treatment is due to blocking fear or blocking the animal's physical ability to freeze. A challenge of conditioning studies more broadly is clarifying whether the effects of interventions are due to fear, memory, or both (Walker & Davis, 1997). In other words, it is often unclear whether attenuated startle or freezing responses result from reduced fear per se or impaired memory of the fear-provoking stimulus resulting from the treatment. 126 P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 Table 1 Experiential characteristics of fear and anxiety. Dimension Fear Anxiety References Emotional valence Temporal focus Duration of arousal Defensive direction Negative Negative For example, Epstein (1972) Presentfocused Phasic (brief) Avoidance (escape) FutureFor example, Tellegen (1982) focused Tonic For example, Davis (1998) (sustained) Approach For example, Epstein (1972); Gray and McNaughton (2000); McNaughton and Corr (2004); Mineka (1979); and Öhman (1993) Diffuse For example, Davis (1998); Lang, Davis, and Öhman (2000); and Macleod and Rutherford (1992) Enhanced For example, Rhudy and Meagher (2000); and Tang and Gibson (2005) Specificity of threat Specific Pain perception Blunted 3.1.1.1. Neurobiology. Despite these limitations, researchers have continued to isolate specific brain areas involved in fear conditioning. Blanchard and Blanchard (1972) implicated the amygdala in fearful responding by observing the failure of rats with amygdalar lesions to freeze when presented with a shock-associated CS. Later studies implicated the lateral nucleus (LA; e.g., LeDoux, Cicchetti, Xagoraris, & Romanski, 1990) and central nucleus (CeA; e.g., LeDoux, Iwata, Cicchetti, & Reis, 1988) of the amygdala. Whereas the LA is implicated only in the acquisition of fear, the CeA is implicated in the acquisition and expression of fear conditioned responses (Wilensky, Schafe, Kristensen, & LeDoux, 2006). In a review of the fear-potentiated startle reflex literature, Davis (2006) highlighted several studies implicating the CeA as the primary circuitry in the acquisition and expression of fear-potentiated startle in rats. Studies have also implicated the CeA in non-human primates (e.g., Kalin, Shelton, & Davidson, 2004). Overall, Davis (2006) concluded that the conditioned fear-potentiated startle reflex has a very short latency, is consistent, and deteriorates rapidly after the removal of the feared stimulus. Fear-conditioning research also implicates prefrontal regions in the extinction of fear. Morgan, Romanski, and LeDoux (1993) demonstrated that lesions to the medial prefrontal cortex (mPFC) exerted no effect on fear conditioning but impaired extinction training in a freezing paradigm. Conversely, Gewirtz, Falls and Davis (1997) found that lesions to the mPFC had no effect on fear conditioning or extinction in a startle paradigm. Although these findings suggest that prefrontal circuits might have a differential involvement in freezing versus startle, Quirk, Russo, Barron, and LeBron (2000) found that ventro-medial prefrontal cortex (vmPFC) lesions exclusively impaired extinction recall. That is, vmPFC lesions did not affect initial extinction learning, but resulted in diminished recall of extinction on subsequent days. In a later study, Sierra-Mercado, Corcoran, Lebrón-Milad, and Quirk (2006) temporarily inactivated the vmPFC by infusing tetrodotoxin (a potent neurotoxin) prior to fear acquisition and prior to extinction. They found that inactivation prior to extinction led to diminished recall of extinction the following day, whereas inactivation prior to fear acquisition had no effect on recall. These findings suggest that the vmPFC may play a role in the maintenance of trait fear by impairing extinction memory for feared stimuli. 3.1.2. State anxiety In the animal literature, state anxiety is often studied in rats using the open field test and the light/dark box. These tasks take advantage of rats' preference for enclosed, familiar spaces and dark environments. When given a choice between a darkened area and a brightly lit area, rodents will consistently choose the dark area (Crawley, 1981). However, anxiolytics reduce the preference for darkness (Onaivi & Martin, 1989). In one study investigating the startle response in a well-lit environment, rodents displayed increased startle responses (Walker & Davis, 1997). Consistent with several theories of anxiety, the light ostensibly represents a diffuse threat and uncertain environment characteristic of anxiety-inducing situations. A benefit of this paradigm over the fear-potentiated startle is that it does not rely on conditioning (Walker & Davis, 1997). Contrary to the extinction displayed in fear-potentiated startle paradigms, repeated exposure to the light without a US does not extinguish elevated responding. Moreover, the rapid dissolution characteristic of the fearpotentiated startle response is not observed in light-enhanced startle. 3.1.2.1. Neurobiology. Although lesions to the CeA diminished startle responses in fear-conditioning paradigms, they did not change the startle response in the light-enhanced paradigm (Walker & Davis, 1997). Conversely, lesions to the bed nucleus of the stria terminalis (BNST) fail to diminish startle responses in fear-conditioning paradigms (Hitchcock & Davis, 1991), but diminish startle in the light-enhanced paradigm (Walker & Davis, 1997). These findings suggest two separable pathways mediating the startle response in fear versus anxiety. However, the CeA and BNST share similar downstream connections (Davis, 2006). In addition to the amygdala, Gray (1982) and Gray and McNaughton (2000) posited that the septum and the hippocampus are central to anxiety. Through several anxiolytic and lesion studies (see Gray, 1976), he demonstrated that anxiolytics and hippocampal lesions produce similar behavioral effects. Moreover, he found that among classical anxiolytics, the common electrophysiological “signature” involved the hippocampal theta rhythm. This rhythm refers to patterns of neuronal firing that produce a slow wave of approximately 5–10 Hz (McNaughton & Gray, 2000). In a review, McNaughton, Kocsis and Hajos (2007) found that manipulation of the theta rhythm is characteristic of all clinically effective anxiolytic agents. They also found that the same was not true for some other drug classes, such as antipsychotics. In support of this finding, studies have found that selective serotonin reuptake inhibitor (SSRI) injections into the hippocampus reduce anxious behaviors in rats (Degroot & Nomikos, 2005). Since Gray and McNaughton (2000)'s work, several studies have supported the role of the hippocampus in anxiety (e.g., Edinger & Frye, 2006). 3.1.3. Trait fear and trait anxiety In addition to the voluminous literature investigating state fear and anxiety in non-human animals, there is a large literature investigating trait anxiety. Conversely, there is little evidence for trait fear (Miller, Garner, & Mench, 2006). However, Aguilar et al. (2003) found trait anxiety (measured using exploration tasks) and fear (measured using avoidance tasks) factors independent of a conditioned fear factor using several behavioral tasks in rats. Trait anxiety in mice and rats is typically studied using animals bred for either high or low levels of anxious behaviors. In non-human primates, these animals are usually selected based on their performance on unconditioned tests. In addition to different levels of trait anxiety, studies have found that different strains of rats and mice are differentially affected by anxiolytic drugs (e.g., Vendruscolo, Takahashi, Brüske, & Ramos, 2003). One limitation to this literature is that the behaviors associated with these tests are often referred to as “anxious/fearful,” reflecting a failure to distinguish these two constructs (e.g., Williamson et al., 2003). With that caveat in mind, Kalisch et al. (2004) found that selectively bred trait anxious rats displayed reduced dorsomedial prefrontal cortex (dmPFC) oxygenation compared with nonanxious rats. De Wit, Yutaka, Balleine, and Dickinson's (2006) findings also support the role of the dmPFC in trait anxiety. They found that rats with intact dmPFC were able to come to a resolution in a goal-conflict paradigm, whereas rats with chemically inactivated dmPFC were not. P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 In one study not using selectively bred rats, Blundell and Adamec (2007) observed trait anxiety-like behaviors following the unprotected exposure of a rat to a cat for 5 min. In this paradigm, the rat is stalked and attacked (bitten and clawed). Following this encounter, rats display enhanced startle (Blundell, Adamec, & Burton, 2005) and anxious behaviors (Cohen et al., 2004) for several weeks, even in the absence of apparent danger cues. Blundell and Adamec (2007) found that the behavioral changes result from changes to NMDA receptors in several amygdala nuclei, the BNST, and the periaqueductal gray. 3.1.4. Implications of non-human animal neuroscience studies Studies investigating the biological substrates of fear and anxiety in non-human animals provide several important implications for our understanding of the trait and state emotions: (1) Autonomic arousal occurs during fear and anxiety states. (2) Two distinct neural pathways appear to mediate fearful and anxious responding: The CeA is the primary circuitry in the expression of fear, whereas the BNST is the primary circuitry in the expression of anxiety. (3) Conditioned fear responses extinguish in the absence of CS/US pairings, whereas unconditioned anxiety responses may not. (4) The fearful response dissipates quickly, whereas anxiety promotes a sustained response. (5) Fear occurs in situations of certain threat, whereas anxiety occurs in response to uncertain situations. (6) Preliminary evidence suggests that the vmPFC is involved in the maintenance of trait fear, whereas the dmPFC is involved in the maintenance of trait anxiety. (7) Overall, these findings support the biological differentiation of fear from anxiety in non-human animals. 3.2. Human studies Although there are several advantages to studying the biological correlates of fear and anxiety in animals, human research also holds several advantages, including the ability to measure subjective levels of fear and anxiety, the ability to manipulate experimental paradigms with verbal instructions, and the ability to measure the effects of implicit as well as explicit stimuli. However, one disadvantage is that humans readily habituate to experimental threat cues (Klorman, 1974). Moreover, it is ethically impermissible to induce state fear comparable to that presumably induced in non-human animal studies. 3.2.1. State fear Fear studies in humans typically involve fear-conditioning paradigms consistent with those in the animal literature (e.g., shocks or air blasts). Additionally, fearful responding is measured following presentation of unconditioned fear-inducing stimuli (such as fearful faces). A variety of techniques is used in these paradigms, including functional magnetic resonance imaging (fMRI), positron emission tomography (PET), fear-potentiated startle, and skin conductance responding (SCR). Fear conditioning in humans consistently increases SCR (e.g., Grillon & Ameli, 2001) and fear-potentiated startle (e.g., Hamm & Vaitl, 1996). Moreover, research demonstrates that, unlike SCR, fear-potentiated startle response is an indicator of arousal and valence, specifically negative valence (Ruiz-Padial & Vila, 2007). Ruiz-Padial and Vila (2007) further found that increased fear-potentiated startle does not require the subjects' awareness of unconditioned fearful stimuli, suggesting that the acquisition of fear may occur as part of implicit (unconscious) processing. Jovanovic et al. (2006) elaborated on this point by finding that startle responses increased to a CS whether or not subjects were aware of it. However, the inhibition of the startle response in the presence of safety cues occurred only in the aware group. In addition to 127 external cues, the human fear-potentiated startle increases when visualizing fearful stimuli (Vrana & Lang, 1990). Although these findings indicate similarities in human and animal fear conditioning, they suggest two implications for the human experience of fear that differ from those found in the animal literature. First, humans are capable of processing environmental fear cues outside of their awareness. Second, internal fear cues can generate human fear reactions. Studies investigating central nervous system responding to fearconditioning paradigms parallel the animal literature. Several studies have found increased blood flow in the amygdala using PET during the acquisition of fear conditioning (e.g., Furmark, Fischer, Wik, Larsson, & Fredrikson, 1997) and increased amygdalar activity using fMRI (e.g., Pine et al., 2001). Supporting Ruiz-Padial and Vila's (2007) findings, studies have found increased amygdalar activity in response to unconditioned fearful stimuli, whether or not the subject was aware of the stimuli (e.g., Whalen et al., 1998). Consistent with the animal literature, several studies have implicated the vmPFC in the recall of fear extinction (see Sotres-Bayon, Cain, & LeDoux, 2006, for a review), whereby activation in the vmPFC is associated positively with extinction recall. Milad, Quinn, Pitman, and Orr (2005) also found that vmPFC thickness, most notably the medial orbitofrontal cortex, was positively associated with extinction recall. In a more recent study, Milad et al. (2007) found that both hippocampal and vmPFC activation were positively associated with extinction learning. 3.2.2. State anxiety Like the state fear literature, the animal literature on state anxiety is broadly corroborated by human studies. In contrast to nocturnal animals, humans experience increased anxiety in response to darkness. In light of this fact, Grillon, Pellowski, Merikangas, and Davis (1997) found preliminary evidence that darkness enhanced the human startle reflex. Nevertheless, they also found that self-reported state anxiety did not correlate with the magnitude of this reflex. Moreover, subjects reported the dark and light conditions as equally unpleasant. One limitation to this study is that participants were presumably healthy undergraduates who volunteered to participate. Therefore, negative findings concerning self-reported anxiety may be due to a restricted range at the lower bounds of trait anxiety, although the authors did not report descriptive statistics for self-reported anxiety. A second limitation is that subjects displayed habituation to both light and dark blocks after the first trial. Despite their limitations, these findings suggest that dark spaces may potentiate human startle. In contrast to fear studies, many studies investigating the neural bases of human state anxiety are equivocal. One primary limitation to these studies is that many use fear-related stimuli, such as fearful faces (e.g., Bishop, Duncan, Brett, & Lawrence, 2004) to elicit anxiety. A secondary limitation is the seemingly conflated interpretation of the data. For example, Straub, Schmidt, Weiss, Mentzel and Miltner (2009) monitored 16 healthy female participants using fMRI while the participants anticipated 4 levels (mild to strong) of electric foot shock. Following the paradigm, participants rated the level of anxiety they experienced on a 9-point Likert-type scale. They found that activation in the pregenual anterior cingulate cortex (ACC) was positive during moderate threat and negative during strong threat. They also found that self-reported anxiety was positively related to dorsal ACC activation during the anticipation of strong threat. The authors interpreted this finding as suggesting that activation of ACC areas is dependent on the level of anxiety and severity of threat, whereby moderate threat is associated with increased attentional avoidance and strong threat with hypervigilance. However, it is unclear whether the authors were measuring self-reported anxiety or self-reported fear, as participants were probably rating their subjective experience of aversive arousal. Moreover, the authors labeled brain activation during a fear-conditioning paradigm as “anxiety.” 128 P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 3.2.3. Trait fear and trait anxiety Some studies have explicitly investigated the physiological and neurobiological substrates of trait fear and anxiety in humans. White and Depue (1999) found that self-reported trait fear was inversely related to pupil dilation in response to the introduction of a norepinephrine (which facilitates global alarm functions) agonist, whereas trait anxiety was unrelated to pupil dilation. Additionally, they found that pupil dilation in response to darkness (an evolutionary threat, as noted earlier) was positively related to trait anxiety but unrelated to trait fear. They interpreted these findings as implying that trait fear and trait anxiety are different emotional dispositions. In terms of central nervous system functioning, several early studies (e.g., Gur et al., 1988; Martinot et al., 1990) found mixed results as to lateralization of fear versus anxiety. Some authors have found that anxiety is right-lateralized (e.g., Gur et al., 1988), whereas others have found that anxiety is bilateral (e.g., Reiman et al., 1989). However, Heller, Nitschke, Etienne, and Miller (1997) suggest that these findings are equivocal because studies often recruited participants from anxiety disordered populations and failed to differentiate whether fear (e.g., specific phobia) or anxiety (e.g., generalized anxiety disorder) was the primary emotion underpinning the anxiety disorder of interest. To address ambiguity regarding the lateralization of trait anxiety, Heller and colleagues administered the trait scale of the State-Trait Anxiety Inventory (STAI; Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983) to an initial pool of 2203 participants. From this pool, they selected 20 highly anxious and 20 comparison participants. The experimental task consisted of measuring EEG while participants listened to 30-second, emotionally aversive narratives. The authors found that greater hemispheric asymmetry, specifically more left hemispheric activity, differentiated anxious from comparison participants. Dien (1999) found that self-reported trait fear correlated with increased right-lateral blood flow in the frontal lobe during an objectspatial recognition task, whereas trait anxiety correlated with increased left-lateral blood flow in the frontal lobes. Funayama, Grillon, Davis, and Phelps (2001) found some support for Dien's findings with a double dissociation of the startle reflex in patients with right or left temporal lobe damage. Patients with right temporal lobe damage reacted appropriately to a task involving uncertainty of threat (i.e. state anxiety), but displayed impaired reaction to visual presentation of threatening stimuli (i.e. state fear). Patients with left temporal lobe damage, in contrast, reacted appropriately to the visual presentation of threatening stimuli (i.e., state fear), but displayed impaired reaction to the uncertainty task (i.e., state anxiety). Moreover, Morinaga et al. (2007) found that right prefrontal brain activity was associated with self-reported fear, but not anxiety, during a fear induction paradigm. Engels et al. (2007) sought to differentiate neural activity associated with high trait fear (which they referred to as “anxious arousal”) versus high trait anxiety (which they referred to as “anxious apprehension”) during an emotional Stroop task. They prescreened a group of 1099 participants and selected 11 scoring low on trait fear but high on trait anxiety, 13 scoring high on trait fear but low on trait anxiety, and 18 scoring low on both trait fear and trait anxiety. Their results indicated that the trait fearful group exhibited more activity in the right inferior temporal gyrus for negative than neutral words, whereas the trait anxious group exhibited more activity in the left inferior frontal gyrus for negative than neutral words. Consistent with Dien (1999), their findings suggest that trait fear is characterized by increased right-lateral blood flow, whereas trait anxiety is characterized by increased left-lateral blood flow. 3.2.4. Implications of human neuroscience studies In summary, human neuroscience studies permit several conclusions: (1) Autonomic arousal occurs during both fear and anxiety states. (2) Preliminary evidence for the lateralization of trait fear and anxiety, whereby fear is right-lateralized and anxiety left-lateralized. (3) Impaired extinction but not acquisition of fear characterizes trait fear, supporting the conceptualization of trait fear as persistent harm avoidance behavior even when it is irrational. (4) Enhanced arousal during the acquisition phase of fearconditioning paradigms characterizes trait anxiety, supporting the conceptualization of trait anxiety as hypervigilance. (5) Evidence suggests that fear and anxious responding can occur inside or outside of awareness, challenging the conceptualization of fear or anxiety as purely cognitively mediated. 4. Operationalizations of fear and anxiety: Consensus and disagreement Clinical and neuroscientific theorists have offered many conceptualizations of fear and anxiety that converge on some points and diverge on others. Although biological similarities and differences do not necessarily imply experiential differences, the findings from this literature suggest subjective differences in the certainty of threat, direction of defensive response, and duration of experience. In defining fear and anxiety, it becomes necessary to focus on trait emotions rather than clinical disorders or syndromes. 4.1. Defining fear By informing theoretical conceptualizations with biological findings, we can discern that state fear has several characteristics (see Table 1). Trait fear appears to result, in part, from an underactive extinction circuit. Therefore, avoidance responses persist even with evidence demonstrating the removal of the objective danger. Although extinction of irrational fears remains possible, the learning curve associated with extinction appears to be higher in trait fearful than in other individuals. In terms of awareness, repeated avoidance reactions to fearful stimuli presumably become automatic over time. Therefore, the trait fearful individual can ostensibly avoid specific threat stimuli outside of awareness and in some cases without recalling the associated danger. In assessing self-reported trait fear, questions addressing avoidant behaviors (that is, the fight, flight, and freeze response) across situations are appropriate. It is not sufficient to ask whether an individual is fearful of specific stimuli without assessing avoidance behaviors. It is likely that individuals cannot discriminate anxiety from fear without appropriate instruction. In other words, individuals may endorse being fearful of aversive stimuli about which they are hypervigilant but do not necessarily avoid. For example, individuals may endorse a fear of flying. However, they may elaborate by stating that they are concerned that something bad will happen when they fly and consequently experience a prolonged state of aversive arousal. In this scenario, individuals are endorsing anxiety surrounding flying (manifested as hypervigilance and hyperarousal during flight), but not fear (avoidance) of flying. Therefore, “How likely are you to avoid flying on an airplane because of fear?” is probably a more valid trait fear item than “How much do you fear flying on airplanes?” 4.2. Defining anxiety Theoretical and biological findings also point to several characteristics of state anxiety (see Table 1).Trait anxiety appears to result largely from a hypersensitive appraisal circuit. Therefore, persistent hypervigilance and prolonged hyperarousal seem to result from overestimating the potential for threat in ambiguous situations. In humans, this sustained state of hyperarousal manifests itself as apprehension, hypervigilance, and rumination. In assessing self-reported trait anxiety, generalized questions addressing apprehension and hypervigilance across situations are probably most valid. Similar to fear questionnaires, it may not be sufficient to ask whether an individual experiences anxiety P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 regarding a specific stimulus or situation. An individual may ascribe anxiety to the aversive arousal experienced before avoiding a specific stimulus (i.e., fear). Therefore, items that assess stimulus non-specific and chronic arousal, such as “I am often worried that something bad will happen,” should be most valid for trait anxiety questions. 5. Diagnostic categories of pathological fear and anxiety: Descriptive psychopathology and behavior-genetic findings The current diagnostic system, the DSM-IV, combines disorders of fear and anxiety under the umbrella category of anxiety disorders. However, several clinical researchers have recently questioned the validity of this DSM-IV classification and proposed alternative models that are aligned more closely with research evidence (e.g., Krueger, 1999; Watson, 2005). Krueger (1999) factor analyzed data from 8098 non-institutionalized participants ranging in age from 15 to 54. To investigate common “comorbidities” (covariations) among mental disorders, Krueger included generalized anxiety disorder (GAD), panic disorder (PD), major depressive disorder (MDD), alcohol dependency, substance dependency, social phobia (SoP), simple phobia (SiP; known as “specific phobia” in DSM-IV), agoraphobia (AG), antisocial personality disorder, and dysthymic disorder (DD) in his analyses. The best fitting model consisted of two correlated lower-order factors, AnxiousMisery (MDD, GAD, and DD) and Fear (SoP, SiP, AG, and PD), that loaded on a higher-order Internalizing Disorders factor. These findings suggest that the DSM-IV category of anxiety disorders is a heterogeneous mix of conditions, some marked primarily by anxiety and others marked primarily by fear. These findings also suggest that at least some mood disorders may primarily be disorders of anxiety or negative emotionality more broadly, not fear. Watson (2005) reviewed the factor-analytic literature on structural models of DSM-IV diagnoses and proposed a revised structural model consisting of three factors. The Emotional Disorders factor comprised Bipolar Disorders, Distress Disorders, and Fear Disorders sub-factors; the Distress Disorders sub-factor comprised of MDD, DD, GAD, and Posttraumatic Stress Disorder; and the Fear Disorders subfactor comprised PD, AG, SoP, and SiP. Consistent with Krueger (1999), Watson proposed separate factors for primarily fear-based versus anxiety-based disorders. Behavioral-genetic studies have similarly pointed to alternative classification models consistent with a distinction between fear- and anxiety-based disorders. Kendler, Prescott, Myers, and Neale (2003) acquired twin diagnostic interview data from more than 5600 individuals in the Virginia Twin Registry using raters blind to cotwin diagnostic status. To increase the number of individuals meeting criteria for PD and GAD, Kendler and colleagues reduced the minimum duration of GAD symptoms from 6 months to 1 month and the minimum number of PD criteria for panic attacks to 2 within 30 min. Their results indicated that MDD and GAD loaded strongly onto an Anxious-Misery sub-factor, whereas animal phobias and situational phobias (e.g., fears of closed spaces, fears of elevators) loaded strongly onto a Fear sub-factor, similar to Watson's (2005) and Krueger's (1999) conceptualizations. Moreover, the two sub-factors again loaded on a higher-order Internalizing Disorders factor. Nevertheless, they found that PD loaded weakly on both the Anxious-Misery and Fear sub-factors, in contrast to Watson (2005) and Krueger (1999). That ambiguity aside, the collective findings point to different genetic underpinnings for primarily anxiety-based and primarily fear-based clinical conditions. 6. Evaluation of the content of self-report fear and anxiety measures Self-report measures are the primary tools for assessing fear and anxiety in studies of anxiety disorders. There are several hundred self- 129 report measures of fear or anxiety in the literature (see Antony, Orsillo, & Roemer, 2001). Many of these measures assess specific components of fear or anxiety (e.g., dental fear), some assess relevant clinical syndromes (e.g., agoraphobia), and others assess related phenomena (e.g., anxiety sensitivity, namely the fear of anxiety). Nevertheless, few measures assess trait fear or anxiety more broadly. Of the trait measures, the most commonly used fear measures are the Fear Questionnaire (FQ; Marks & Matthews, 1979) and the Fear Survey Schedule (FSS; Wolpe & Lang, 1964), and the most commonly used anxiety measures are the Manifest Anxiety Scale (MAS; Taylor, 1953) and the STAI — Trait Scale. One or more of these measures were used in 84% of the studies in the meta-analysis to be reported later. 6.1. Self-report measures of trait fear 6.1.1. Phobic fear measures 6.1.1.1. FQ. The FQ is a 24-item self-report measure designed to assess phobic fears. The first 17 questions assess the specific fears on a 9-point Likert-type scale, with values ranging from 0 (would not avoid it) to 8 (always avoid it). Item content ranges from fears of public transportation to the sight of blood. The last 6 questions are related to anxiety and distress on a 9-point Likert-type scale, with values ranging from 0 (hardly at all) to 8 (very severely troublesome). Their item content ranges from depression to dissociation. The FQ has demonstrated adequate internal consistency, with a range of 0.71 to 0.83 across studies (Oei, Moylan, & Evans, 1991). It also demonstrated a 1-week test–retest reliability of 0.82 (Marks & Matthews, 1979); however, the elapsed time between tests was brief so subjects may have recalled their initial responses. Factor analytic studies of the fear items on the FQ typically find a three factor solution (physical, social, and agoraphobia; Oei et al., 1991), although a few authors have not confirmed this structure (Trull & Hillerbrand, 1990). Moreover, these studies do not support a unitary fear factor. Studies have also demonstrated that the FQ is sensitive to changes in fear following treatment (e.g., Marks & Matthews, 1979). Within the present meta-analysis, all studies limited their analyses to the 17 avoidance items. By the definition of trait fear derived earlier, the first 17 items of the FQ are appropriate for assessing fear across a variety of situations. However, at least some of the FQ content may be outdated or culturally specific. For example, item 5 is “Traveling along by bus or coach.” Although many fears appear to be timeless (e.g., sight of blood), the content of the FQ has remained the same for nearly 30 years and may not accurately reflect common phobic fears in today's society (e.g., radiation, computers, driving on freeways). 6.1.1.2. FSS. There are several versions of the FSS; the most popular are the FSS-II (Geer, 1965) and the FSS-III (Wolpe & Lang, 1964). The FSSII is a 51-item scale developed for research, whereas the FSS-III is a 72item scale developed for clinical settings (Antony et al., 2001). Questions on the FSS-II and FSS-III assess fears on a 7-point Likerttype scale from 0 (None) to 6 (Terror). In contrast to the FQ, the FSS requires subjects to rate the degree of fear they feel towards a broad range of potentially threatening stimuli. The FSS-II and III have demonstrated adequate internal consistency (α N 0.90; Arrindell, 1980). In terms of construct validity, studies have demonstrated a positive relationship between scores on the FSSIII and the magnitude of fear-potentiated startle to fearful stimuli (Cook, Davis, Hawk, Spence, & Gautier, 1992). Factor analytic studies of the FSS-III have yielded mixed results (e.g., Arrindell, 1980; Bates, 1971). However, two factors, blood/physical injury and social, are consistently found across studies. Similar to the FQ, these studies do not support a unitary fear factor. Unlike the FQ, the FSS may not assess fears with adequate validity for two reasons. First, the Likert-type scale does not assess avoidance; second, some of the items (e.g., life after death and God) assess abstract, 130 P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 diffuse, and unavoidable threats akin to anxiety-provoking stimuli. Among the items, 43/51 (84%) assess potentially fear-inducing (rather than strictly anxiety-inducing) stimuli. Using the FQ and FSS-III, Stravynski, Basoglu, Marks, Sengun, and Marks (1995) found evidence for the relative independence for the diagnoses of social, agoraphobic, and specific phobias using the FQ and FSS-III. There are no published studies suggesting that a unitary fear factor provides the best fit to selfreported fear measures. There are also no published studies of the correlations among FQ and FSS-III total and factor scores. 6.1.2. Harm avoidance measures 6.1.2.1. Activity Preference Questionnaire (APQ; Lykken et al., 1973). The APQ is a 74-item forced-choice measure designed to assess fear. Each item consists of two choices, one of which is unpleasant primarily because it is frightening or embarrassing (e.g., “Sleeping out on a camping trip in an area where rattlesnakes have been reported,” “Having to blow your nose while in a group of strangers”) and the other of which is unpleasant primarily because it is burdensome or onerous (e.g., “Sitting through a two-hour concert of bad music,” “Having a gabby old woman sit down next to you on the bus”). Respondents are asked which of the two activities within each pair they would prefer. Within each item, the two choices were matched for social desirability by a panel of judges using a modified Thurstone scaling procedure. The internal consistency of the APQ ranges from 0.82 to 0.86 (Lykken et al., 1973). 6.1.2.2. MPQ — Harm avoidance Scale (HA). The MPQ-HA scale, which was modeled largely after the APQ (Lykken et al., 1973), has been used in a number of studies as a measure of fearfulness. This scale consists of 28 items; many of the items are in forced-choice format and juxtapose a potentially fear-inducing activity (e.g., “Being chosen as the ‘target’ for a knife-throwing act”) with another that is less so (e.g., “Being sick to one's stomach for 24 hours”). Other items ask people how likely they would be to fear or avoid a potentially scary stimulus. A high score on HA indicates a preference for avoiding potentially hazardous situations. Like its “parent” measure, the APQ, the HA scale has adequate internal consistency; for example, across four samples (some including college students, others members of the community), Tellegen and Waller (2007) reported Cronbach's alphas ranging from 0.82 to 0.84. 6.1.2.3. Psychopathic Personality Inventory — Fearlessness Scale (PPI-F; Lilienfeld & Andrews, 1996). The PPI-F scale was designed to assess fearlessness or (reversed HA), considered by many to be a core attribute of psychopathy (Lykken, 1995). This scale consists of 19 items in a 1–4 Likert-type format and, like other PPI scales, was constructed using an iterative approach in which the results of successive exploratory factor analyses informed construct reformulation and item revision (Lilienfeld & Andrews, 1996). Each item assesses the relative absence of fear, typically in one or more specific situations or settings (e.g., “I would find the job of movie stunt person exciting”). Across two undergraduate samples, Lilienfeld and Andrews (1996) reported internal consistencies (Cronbach's alphas) of 0.86 and 0.88 for the PPI-F scale. Correlations between this scale and trait anxiety measures were reflected in direction in the analyses reported here given that this scale assesses reversed fear/HA. Among the items on the MAS, 4/50 (8%) are explicit fear items (e.g., “I do not have as many fears as my friends”) and 5/50 (10%) are related ambiguously to anxiety (e.g., “I cry easily”). Therefore, 41/50 (82%) appear to be explicit anxiety questions. 6.1.3.2. STAI-T. The STAI-T is a 20-item scale designed to assess pervasive feelings of anxiety. Items are rated by respondents on a 4-point Likert-type scale. The STAI-T has demonstrated acceptable internal consistency (α N 0.85) and 1-month test/retest reliability (r N 0.70) in adolescent, healthy adults, and military samples (Spielberger et al., 1983). Factor analytic studies (e.g., Bieling, Antony, & Swinson, 1998) suggest that a two-factor model, consisting of depression and anxiety, best fits the data. Bieling and colleagues found that the depression factor correlated more strongly with measures of depression than anxiety, whereas the anxiety factor correlated more strongly with measures of anxiety than depression. This finding and others suggest that the STAI-T is heterogeneous in content and indexes construct-irrelevant variance (Messick, 1995) stemming from mood disturbance. Among the items on the STAI-T, 5/20 (25%) are explicit depression items (e.g., “I wish I could be as happy as others seem to be”) and 1/20 (5%) are ambiguously related to anxiety (e.g., “I have disturbing thoughts”). Therefore, 14/20 (70%) are explicit anxiety questions. Despite the differences in format and item content, correlations between the STAI-T and MAS are high, ranging from 0.72 to 0.92 (e.g., Watson & Clark, 1984). 7. Meta-analysis of self-report fear and anxiety measures Our meta-analysis sought to evaluate the relationship between scores on frequently used measures of self-reported trait fear and trait anxiety. Based on the clinical and neurobiological literatures, we generated the following hypotheses. 7.1. Hypothesis 1: Primary meta-analysis As the preponderance of evidence from these literatures suggests that trait fear and trait anxiety are separable phenomena, we hypothesize that the relationship between self-reported trait fear and trait anxiety will be low or at best moderate. Although three author teams who conducted narrative reviews (Perkins et al., 2007; Watson & Clark, 1984; White & Depue, 1999) have reported that the correlations between trait fear and trait anxiety appear to be low, none has examined this association meta-analytically. 7.2. Hypothesis 2: Self-report measure as moderator We also investigate the moderating effect of self-report measure. From our evaluation of the content of commonly used trait fear and anxiety measures, we found that some measures contain less unrelated or overlapping items than other measures. Therefore, we hypothesize that the relationship between fear and anxiety will be significantly weaker in the FQ versus the FSS as well as the MAS versus the STAI-T. Moreover, we hypothesize that the relations between trait anxiety and trait fear assessed using HA measures (e.g., MPQ-HA) will be significantly weaker than with trait fear assessed using specific fear measures (e.g., FQ). 6.1.3. Self-report measures of trait anxiety 6.1.3.1. MAS. The MAS is a 50-item true/false scale of trait anxiety derived from the Minnesota Multiphasic Personality Inventory (MMPI). The internal consistency of the MAS is typically high (α N 0.9; King & Campbell, 1986). Using MAS data, Buss (1962) found two primary anxiety factors, physiological hyperreactivity and subjective feelings of apprehension. Fenz and Epstein (1965), however, found that a unitary anxiety factor accounted for nearly 75% of the variance in the scale. 7.3. Exploratory analyses: Age, gender, and psychiatric status as moderators In addition to testing these two hypotheses, we conduct exploratory analyses investigating whether age, gender, and psychiatric population moderate the relation between trait fear and trait anxiety. Because we do not have access to the ages of participants in the studies, we investigate age as a categorical moderator with child P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 Table 2 Phobic fear and anxiety scale correlations in child samples. 131 Table 4 Harm avoidance and anxiety scale correlations in adult samples. Study Measures N r ESzr Study Byrne (2000) King, Gullone, and Ollendick (1992) Muris et al. (2003) Sample 1 Muris et al. (2003) Sample 2 Muris et al. (1998) Ollendick (1983) Sample 1 Ollendick (1983) Sample 2 Peleg-Popko and Dar (2001) Scherer and Nakamura (1968) Schmidt and Mallott (2005) Shore and Rapport (1998) Silverman, Fliesig, Rabian, and Peterson (1991) Walters (2001) Wilson and Hayward (2006) FSS-C; STAI-C FSS-C-R; RCMAS KFS; STAI-T KFS; STAI-T FSS-C; SCRAD FSS-C; STAI-C FSS-C; STAI-C CFI; CSAI FSS-II; MAS-C FNES; STAI-T FSS-R; MAS-C FSS-C; STAI-T 300 1524 189 163 120 99 118 108 99 405 385 81 0.50⁎ 0.53⁎ 0.29⁎ 0.59⁎ 0.44⁎ 0.51⁎ 0.46⁎ 0.54⁎ 0.49⁎ 0.33⁎ 0.31⁎ 0.57⁎ 0.55 0.59 0.30 0.68 0.47 0.56 0.50 0.60 0.54 0.34 0.32 0.65 Blankstein (1976) Church (1994) Helfritz and Stanford (2006) Hodges and Felling (1970) Lilienfeld and Andrews (1996) Lykken et al. (1973) Newman and Schmidt (1998) Poythress et al. (n.d.) Schmidt and Newman (1999) Sellbom and Ben-Porath (2005) FQ-C; STAI-C FQ; STAI-T 149 2246 0.50⁎ 0.31⁎ 0.55 0.32 Note. ESzr = Fisher's Zr-transform; FNES = Fear of Negative Evaluation Scale. FQ = Fear Questionnaire. FSS = Fear Survey Schedule. FSS-C = Fear Survey Scale — Child Version. MAS-C = Manifest Anxiety Scale — Child Version. RCMAS = Revised Child Manifest Anxiety Scale. STAI-C = State-Trait Anxiety Inventory — Child Scale. STAI-T = StateTrait Anxiety Inventory — Trait Scale. ⁎ p b 0.05. and adult categories. As the anxiety disorders of PD (e.g., Oei, Evans, & Crook, 1990) and GAD (e.g., Oei et al., 1991) are not necessarily associated with trait anxiety and trait fear, respectively, we conduct exploratory analyses to investigate whether the relationship is weaker in clinical than non-clinical samples. As we do not have specific diagnostic information for most participants, we investigate psychiatric versus community samples as a categorical moderator. Table 3 Phobic fear and anxiety scale correlations in adult samples. Study Measures N Bates (1971) Carleton, Collimore, and Asmundson (2007) Carsrud and Carsrud (1979) Dien (1999) Geer (1965) Grossberg and Wilson (1965) Hagopian and Ollendick (1996) Hersen (1971) Kent and Keohane (2001) Kilpatrick and McLeod (1973) Kilpatrick, Sutker, Roitzsch, and Mason (1975) Kocovski and Endler (2000) Kogan and Edelstein (2004) Lang and Lazovik (1963) Lerner and Keltner (2001) Study 1 ⁎Lerner and Keltner (2001) Study 2 ⁎Manosevitz and Lanyon (1965) ⁎Mattick and Clarke (1998) Mellon (2000) Perkins et al. (2007) Sample 1 Perkins et al. (2007) Sample 2 Schroeder and Craine (1971) Stanley, Beck, and Zebb (1996) Sample 1 Stanley et al. (1996) Sample 2 Weeks et al. (2005) FSS-III; MMPI-A BFNE; SIAS 41 322 r ESzr FSS; MAAC FSS; STAI-T FSS-II; MAS FSS-III; MAS FQ; STAI-T FSS-III; MAS FNES; HADS-A FSS-III-M; STAI-T FSS-III-F; STAI-T 99 − 0.07 − 0.07 63 0.40⁎ 0.42 205 0.42⁎ 0.45 505 0.46⁎ 0.50 61 0.38⁎ 0.40 351 0.46⁎ 0.50 141 0.48⁎ 0.52 36 0.52⁎ 0.58 211 0.39⁎ 0.41 FNES; EMAS-S FSS-II; BAI FSS; MAS FSS-II; STAI-T FSS-II; STAI-T FSS-III; MAS FQ-SP; STAI-T FSS-GV; MMPI-A FSS-III; STAI-T FSS-III; STAI-T FSS-III; MAS FQ; STAI-T FQ; STAI-T BFNE; SIAS 174 114 13 75 601 46 42 696 141 101 107 50 94 148 0.57⁎ 0.64⁎ 0.41⁎ 0.40⁎ 0.81⁎ 0.54⁎ 0.57⁎ 0.27 0.51⁎ 0.49⁎ 0.21 0.34⁎ 0.56⁎ 0.34⁎ 0.43⁎ 0.38⁎ 0.65 0.76 0.44 0.42 1.13 0.60 0.65 0.28 0.56 0.54 0.21 0.35 0.63 0.35 0.46 0.40 Note. ESzr = Fisher's Zr-transform; BAI = Beck Anxiety Inventory. BFNE = Brief Fear of Negative Evaluation Scale. EMAS = Endler Multidimensional Anxiety Scale. FNES = Fear of Negative Evaluation Scale. FQ = Fear Questionnaire. FSS = Fear Survey Schedule. FSS-GV = Fear Survey Schedule — Greek Version. HADS-A = Hospital Anxiety and Depression Scale — Anxiety. MAAC = Multiple Affective Adjective Scale. MAS = Manifest Anxiety Scale. MMPI = Minnesota Multiphasic Personality Inventory. SIAS = Social Interaction Anxiety Scale. STAI-T = State-Trait Anxiety Inventory — Trait Scale. ⁎ p b 0.05. Measures APQ; STAI-T MPQ-HA; NEO-A PPI-F; PAI-A SSQ-P; STAI-T PPI-F; MPQ-SR APQ; STAI-T MPQ-HA; WAS MPQ-HA; STAI-T MPQ-HA; WAS MPQ-HA; MMPI-2 RCP Uzieblo, Verscuerea, and Crombez PPI-F; STAI-T (2007) Waller, Lilienfeld, Tellegen, and MPQ-HA; TPQ Lykken (1991) N r 101 674 39 228 110 25 207 1473 217 985 0.05 0.05 0.20⁎ 0.20 0.23 0.23 0.02 0.02 0.06 0.06 − 0.01 − 0.01 − 0.22⁎ − 0.22 − 0.05 − 0.05 − 0.15 − 0.15 0.04 0.04 ESzr 120 0.14 0.14 1236 0.22⁎ 0.22 Note. ESzr = Fisher's Zr-transform; APQ = Activities Preference Questionnaire. MMPI-2 RCP = Minnesota Multiphasic Personality Inventory — Restructured Clinical Scale Psychasthenia. MPQ-HA = Multidimensional Personality Questionnaire — Harm avoidance Scale. MPQ-SR = Multidimensional Personality Questionnaire — Stress Reaction Scale. PAI-A = Personality Assessment Inventory — Anxiety Scale. PPI-F = Psychopathic Personality Inventory, Fearlessness Scale, reverse scored. SSQ-P = Stressful Situations Questionnaire — Physical Danger Scale. STAI-T = State-Trait Anxiety Inventory — Trait Scale. ⁎ p b 0.05. 8. Method 8.1. Literature search and selection criteria We identified relevant studies primarily through a computerized database search of journal articles using Psycinfo and Medline with the constraints of human studies, peer-reviewed journals, written in the English language using the key words anxiety, anxious, fear, harm avoidance, neurotic, panic, phobia, phobic, and self-report. If the title of a paper included the search terms fear and anxiety, the entry was automatically included in an initial pool of potential studies. In secondary searches, we used the names of commonly used self-report measures of fear and anxiety as key words, including Fear Questionnaire, Fear Survey Schedule, Manifest Anxiety Scale, Multidimensional Personality Questionnaire, Psychopathic Personality Inventory, and State-Trait Anxiety Inventory. In addition to using databases, we inspected reference lists from retrieved articles and books to generate potential studies. Finally, as the relationships of interest were often buried within studies, we perused issues of a specialized anxiety journal, the Journal of Anxiety Disorders, from the publication dates 1987 (the first year of publication) through 2009. Studies reporting within-measure correlations among subscales (e.g., MPQ-Stress Reaction and MPQ-Harm Avoidance, PPI-Stress Immunity and PPI-Fearlessness) were excluded from the meta- Table 5 Gender and correlations between phobic fear and anxiety scales. Study Byrne (2000) Geer (1965) Grossberg and Wilson (1965) Gullone and King (1992) Hersen (1971) Ollendick (1983) 1* Ollendick (1983) 2* Scherer and Nakamura (1968) Measures FSS-C; STAI-C FSS; MAS FSS; MAS FSS-C; STAI-C FSS-III; MAS FSS-C; STAI-C FSS-C; STAI-C FSS-II; MAS-C Male Female N r ESzr N r ESzr 150 150 203 207 160 57 51 59 0.49 0.39 0.42 0.38 0.42 0.44 0.32 0.41 0.54 0.41 0.45 0.40 0.45 0.47 0.33 0.44 150 55 302 225 191 42 67 40 0.50 0.55 0.45 0.31 0.52 0.56 0.50 0.52 0.55 0.62 0.48 0.32 0.58 0.63 0.55 0.58 Note. ESzr = Fisher's Zr-transform; FSS = Fear Survey Schedule. FSS-C = Fear Survey Scale — Child Version. MAS = Manifest Anxiety Scale. MAS-C = Manifest Anxiety Scale — Child Version. STAI-C = State-Trait Anxiety Inventory — Child Version. All correlations reached significance at p b 0.05. 132 P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 analysis. Because these subscales were derived from factor-analytically developed measures that were intended to yield largely independent dimensions, the inclusion of these studies could artificially weaken the relationship between trait fear and trait anxiety (cf., White & Depue, 1999, who included these correlations in their narrative review). In the case of the MPQ, its 11 lower-order scales were designed to be largely independent (Tellegen & Waller, 2007). In the case of the PPI, items that loaded highly on more than one lower-order factor were discarded in the process of subscale development (Lilienfeld, 1990). Hence, our analyses focused exclusively on correlations between fear and anxiety subscales derived from different omnibus measures of personality or psychopathology. These analyses represent a “stringent” test of our central hypotheses, as they exclude numerous studies in which trait fear and trait anxiety correlations are likely to be spuriously low. 8.1.1. Inclusion criteria The following criteria were used to select studies for inclusion in the meta-analysis: 1. The study was based on data published in a peer-reviewed, English journal through December 2009. 2. The study included self-report measures of trait fear and trait anxiety (i.e., studies using measures of a specific fear or anxiety were excluded). 3. The study reported data that allowed the computation of an effect size of the relation between the measures. In several cases where these data were not reported, we contacted authors for their raw data. 4. The study was based on correlations between trait fear and trait anxiety measures derived from different broadband measures. 8.1.2. Search results These search strategies generated a list of approximately 800 articles. These articles were retrieved and examined for inclusion criteria. Only 55 of the 800 articles reported association between selfreported fear and anxiety. Of the 55 articles, we excluded 9 from the meta-analysis for using measures tailored towards specific, rather than trait, fear or anxiety. An additional dataset containing the variables of interest (which have formed the basis for several publications, but in which the intercorrelations between trait fear and trait anxiety were not reported) was identified by correspondence with the authors (Poythress et. al, unpublished). fear measures: FQ (5 studies, n = 2493), FSS (24 studies, n = 4607), and BFNE (5 studies, n = 1190). Second, we compared the phobic fear (FQ, FSS, BFNE; 34 studies, n = 8290) measures' relations with anxiety measures, on the one hand, with those of the HA (PPI-F, MPQ-HA, APQ; 13 studies, n = 6765) measures, on the other. To account for the possible effect of anxiety measure, we conducted a secondary analysis investigating relations between the FQ (4 studies, n = 2451) and FSS (12 studies, n = 2037) with the STAI-T. In the second analysis, we compared two anxiety measures: STAI-T (11 studies, n = 1829) and MAS (7 studies, n = 1305). The FSS was the only fear measure used in the MAS studies, so we only used the 11 studies correlating the STAI with the FSS for the analysis to minimize systematic error. 8.2.1. Computing effect sizes In the current meta-analysis, effect sizes index the relationship between self-report scores on measures of anxiety with those of selfreport measures of fear. Within studies, the effect sizes were reported as Pearson correlations. Effect sizes for individual studies were estimated using the unbiased estimator Fisher's Zr-transform (Hedges & Olkin, 1985), defined as ESzr = 0.5 loge [1 + r/1 − r]. We used Microsoft Excel 2003 to compute all effect-size estimates. 8.2.2. Effect-size aggregation The grand mean, as well as means for each level of a given categorical moderator, was computed by using an estimate of the inverse of each estimates sampling variance (Hedges & Olkin, 1985). This weighting scheme accords greater weight to more precise effectsize estimates (Matt & Cook, 1994). Upon averaging the weighted effect sizes, the overall effect size was converted back into a Pearson correlation for clarity of presentation. 8.2.3. Categorical moderator analyses We compared weighted mean effect sizes from different levels of the same moderator with the Q test statistic. The Q statistic is a between-group homogeneity test, analogous to a pairwise comparison with two categories, derived from formulae in Hedges and Olkin (1985), and defined as Q = Σwi (ESi − ES)2 (Lipsey & Wilson, 2001). The Q statistic is distributed as a chi-square with k − 1 degrees of freedom, where k is the number of groupings. If the Q statistic exceeds the corresponding chi-square critical value, one can infer that the effect differs across groups (e.g., the relation is different in male than in female). 9. Results 8.2. Data analytic strategy 9.1. Hypothesis 1: Primary meta-analysis In total, this search strategy yielded 47 articles suitable for the current meta-analysis (see Tables 2, 3, and 4). Of these articles, five included more than one sample. For Hypothesis 1, we aggregated 52 effect sizes (N = 17,620) for the meta-analysis. In exploratory analyses, these studies were further evaluated for use in categorical moderator analyses relevant to age, gender, and population. For the analysis of age, 12 studies (n = 4340) reported the relation in nonpsychiatric child samples and 19 studies (n = 3,628) in nonpsychiatric adults samples. Because no studies reported the relation in psychiatric child samples, we excluded psychiatric adult samples from this analysis. For the analysis of gender, eight studies reported the relation in male (n = 1037) and female (n = 1072) separately. For the analysis of population type, 7 studies (n = 921) reported the relation in psychiatric adult populations and 19 studies (n = 3628) reported the relation in non-psychiatric (and non-forensic) adult populations. For Hypothesis 2, the moderator analyses investigated whether the relationships between widely used measures of trait fear and trait anxiety differed significantly across measures. In the first analysis, we compared the relations between trait anxiety and 3 different phobic We computed 52 effect sizes from the studies in the target literature. Of these effect sizes, 39 (75%) were significantly positive, 12 (23%) were nonsignificant, and 1 (2%) was significantly negative. Tables 1, 2, and 3 display effect sizes and relevant study characteristics for each study included in the analysis. Results of the main metaanalysis are consistent with Hypothesis 1 and indicated a moderate relationship (r = 0.32) between trait fear and trait anxiety scales, with correlations ranging from − 0.22 to 0.81. Fail-safe N calculations (Rosenthal, 1979) indicated that an additional 114 studies reporting null results would be required to reduce the overall correlation to 0.10, and 1612 studies to reduce the overall correlation to 0.01. 9.2. Hypothesis 2: Self-report measure as a moderator 9.2.1. Fear measure as a moderator We computed the relations between trait anxiety and 3 measures of phobic fear using 34 effect sizes (5 BFNE samples, 5 FQ samples, and 24 FSS samples). Consistent with Hypothesis 2, our results indicated that the relation with trait anxiety is significantly different P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 (Q = 43.32, df = 2, p b 0.05) between the BFNE (r = 0.45), FQ (r = 0.32), and FSS (r = 0.45). In subsidiary analyses using only one measure of anxiety, 16 effect sizes (4 FQ and 12 FSS) were computed measuring the relation with the STAI-T. Consistent with Hypothesis 2, our results indicated that the relation with trait anxiety is significantly weaker (Q = 39.45, df = 1, p b 0.05) for the FQ (r = 0.32) than for the FSS (r = 0.48). We also tested whether the relations differed between HA measures (6 MPQ-HA, 3 PPI-F, 2 APQ, and 1 Stressful Situations Questionnaire — Physical Danger Scale (Hodges & Felling, 1970) sample) and physical phobic fear measures (FQ and FSS) with trait anxiety. As HA measures were only assessed in adult samples, we only used phobic fear studies in adults (4 FQ and 18 FSS) as a comparison. Consistent with Hypothesis 2, our results indicated that the relation between trait anxiety and HA measures (r = 0.14) is significantly weaker (Q = 308.29, df = 1, p b 0.05) than the relation between trait anxiety and phobic fear measures (r = 0.46). 9.2.2. Anxiety measure as a moderator We used 19 effect sizes (12 STAI-T samples and 7 MAS samples) to test whether the relations with trait fear differed across anxiety measures. Contrary to Hypothesis 2, our results indicated that relations with the FSS are not significantly different (Q = 0.22, df = 1, p = ns) in the STAI-T (r = 0.48) compared with the MAS (r = 0.46). 9.3. Exploratory analyses: Age, gender, and psychiatric status as moderators 9.3.1. Age as a moderator Thirty-one effect sizes (12 child samples and 19 adult samples) were computed from the studies in the target literature. Our results indicated that the correlation between fear and anxiety measures is not significantly different (Q = 2.38, df = 1, p N 0.05) in adults (r = 0.44) and children (r = 0.42). 9.3.2. Gender as a moderator Sixteen effect sizes (8 male and 8 female) were computed to test whether gender moderated the association between trait fear and anxiety (see Table 5). Our results indicated that the relationship is nonsignificantly different (Q = 1.50, df = 1, p = ns) in male (r = 0.41) and female (r = 0.46). 9.3.3. Psychiatric status as a moderator To test the moderating influence of psychiatric status on the relation between trait fear and anxiety, 26 effect sizes (7 psychiatric and 19 non-psychiatric samples) were computed. Our results indicated that the relation is nonsignificantly different (Q = 0.16, df = 1, p = ns) between psychiatric (r = 0.45) and non-psychiatric (r = 0.44) populations. 9.4. Summary of our meta-analytic findings In summary, the overall relation between measures of trait fear with measures of trait anxiety was moderate. The association was significantly weaker among measures of HA than measures of phobic fear. Among phobic fear measures, the relation of trait anxiety with the FQ was significantly weaker than with the FSS. The relation between measures of trait fear and trait anxiety was not moderated by age, gender, or psychiatric status. There were several limitations to our meta-analysis. First, it is unlikely that the search strategy located all relevant studies. Although the strategy was intensive, correlations of interest were often buried within studies as secondary analyses. Therefore, even fairly comprehensive database searches are likely to overlook pertinent studies. Second, the effect sizes were derived from presumably flawed measures. Although error variance is inherent in all self-report 133 measures, a portion of the error variances in these measures was probably systematic due to their flawed conceptualizations of trait fear and anxiety. Third, the data collected for the meta-analysis were not sufficient for testing the potential moderating effect of subtype of phobic fear (e.g., social versus physical injury) on the relation between trait fear and trait anxiety. 10. Discussion and implications Our review points to three broad implications regarding the relation between trait fear and trait anxiety. First, a wealth of research across scientific disciplines suggests that trait fear and trait anxiety reflect different emotions with separable correlates. Second, these literatures suggest that avoidance behaviors across several situations characterize trait fear, whereas sustained hypervigilance and prolonged hyperarousal while approaching several situations characterizes trait anxiety. These literatures dovetail with our finding that the relation between trait fear and trait anxiety was especially weak when the trait fear measures emphasized avoidance (viz., HA). Third, results from the meta-analysis suggest that the subjective experiences of trait fear and trait anxiety are largely independent, although probably at least somewhat overlapping, phenomena. Although the differentiation and operationalization of trait fear from trait anxiety remains a contentious topic among clinical theorists, the preponderance of research evidence we have reviewed demonstrates that the two emotions are separable and derived largely from different etiological substrates. First, the neurobiological literature suggests that state fear and state anxiety are differentiated primarily by activation in the CeA and BNST, respectively. More provisionally, this literature also suggests that trait anxiety is characterized by increased left frontal hemispheric activity, whereas trait fear is characterized by increased right hemispheric activity. Consistent with the neurobiological findings, diagnostic and behaviorgenetic studies have found that disorders characterized by trait fear versus trait anxiety load on separable factors and differ in their genetic underpinnings. The neurobiological literature further supports a behavioral distinction between state fear and state anxiety. Although both emotional states are experienced as aversive and motivate behavior, they push behaviors in opposite directions. State fear is an aversive emotional state during which an organism is motivated to escape a specific and imminent threat. The characteristics of state fear include short-lived arousal that quickly dissipates after the threat is avoided. Trait fear, therefore, is the persistent and pervasive experience of state fear across situations. In colloquial terms, a “scaredy cat” is a vivid description of a trait fearful individual, who avoids taking risks that most perceive as relatively benign but does not appear distressed when these risks are absent. State anxiety, in contrast, is an aversive emotional state that occurs while an organism approaches an ambiguous and uncertain threat. The behavioral characteristics of state anxiety include sustained hypervigilance and hyperarousal, which persist even after the potential for threat is removed. Trait anxiety is therefore the persistent and pervasive experience of state anxiety across situations. In colloquial terms, a “worry wart” is a vivid description of a trait anxious individual, who does not avoid situations but remains persistently hypervigilant for potential threats (Barlow, 2002). Our meta-analysis suggests that the subjective experiences of trait fear and trait anxiety are distinguishable. Nevertheless, one major limitation to this meta-analysis was the content overlap between selfreport measures of trait fear and trait anxiety. Our finding from the investigation of self-report measure as a categorical moderator suggests that the relation is significantly weaker with reduced content overlap. This finding suggests that a refinement of self-report measures of fear and anxiety to reduce cross-contamination of constructs and construct-irrelevant variance (Messick, 1995) is 134 P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 warranted. Specifically, self-report measures of trait fear should emphasize freezing and avoidance behaviors aimed at an array of specific threats, whereas self-report trait anxiety should emphasize hypervigilance, uncertainty, and hyperarousal. In this respect, measures of HA, which our analyses showed to be the most independent of trait anxiety, are presumably more valid for detecting trait fear than are measures of phobic fear, which our analyses showed to be moderately correlated with trait anxiety. Moreover, these measures should directly assess the degree that an individual displays a propensity for these traits, rather than measuring a clinical syndrome (e.g., the inclusion of depression items on the STAI-T). 10.1. Implications for diagnostic classification systems Our literature review and meta-analysis lend some support to Epstein's (1972) clinical conceptualizations of trait fear and anxiety as separable phenomena with differing etiologies, whereas they call into question Beck and Emery's (2005) clinical conceptualization as slightly different aspects of a single underlying phenomenon. Although data from the meta-analysis do not provide unambiguous support for Tellegen's (1985) assertion that trait fear and trait anxiety are virtually orthogonal constructs, our moderator analyses suggested that inappropriate item overlap across scales inflated the relationship. Therefore, Tellegen's position cannot be falsified based on the present results. Moreover, our moderator analyses suggested that, consistent with Tellegen's position, measures of trait anxiety and HA are at best only when weakly correlated. Our results are also consistent with recent analyses suggesting that disorders of fear and disorders of anxiety warrant classification in separate higher-order categories. Results from our meta-analysis point to the need to devise consistent operationalizations of trait fear and trait anxiety in the personality and clinical literatures, and to develop self-report measures consistent with these operationalizations. Because our review indicates that trait fear and trait anxiety are largely independent, it supports Krueger's (1999) and Watson's (2005) contention that the current diagnostic category of anxiety disorders is not justified on empiric grounds. Moreover, it suggests that studies investigating levels of trait fear and anxiety characteristic of mental disorders is warranted. With appropriate assessment measures, these empirical questions would be testable and clearer conceptualizations of the role of trait fear and anxiety in mental disorders would be possible. 10.2. Clinical implications Anxiety disorders, along with impulse control disorders, are among the most commonly diagnosed mental disorders (DuPont et al., 1996). Given the gravity and prevalence of these disorders, research efforts investigating their pathogenesis, differential characteristics, and treatment are numerous. Although many of these studies use measures designed to assess symptoms rather than underlying personality features, there is sufficient evidence for the differentiation of the two emotion systems to suggest that including theoretically and empirically sound measures of trait fear and anxiety in these studies is warranted. Moreover, including measures of trait fear and anxiety in treatment outcome studies of anxiety disorders may help us better understand the mechanisms of change in psychotherapy (e.g., do changes in trait anxiety precede changes in trait fear or vice versa?). 10.3. Implications for neuroscience These findings also bear important implications for neuroscience research related to the treatment of anxiety disorders. When using selfreport measures based on questionable conceptualizations of trait fear and anxiety to study the neurobiological underpinnings of these emotional dispositions, research is unlikely to yield replicable findings across measures and studies. As Heller's et al. (1997) findings suggest, the conflation of the constructs can produce ambiguity in the literature on cerebral laterality and probably other psychophysiological phenomena. Therefore, a refined operationalization of trait fear and trait anxiety, along with construct measures deduced from these definitions, may better elucidate the neural underpinnings of these trait emotions. 10.4. General summary and future directions Overall findings from our review and meta-analysis suggest that trait fear and trait anxiety are separable constructs. Nevertheless, many widely used self-report measures used to assess these constructs are largely inadequate in light of the existing clinical and neuroscience literatures. Therefore, studies aimed at developing more construct valid – and less construct contaminated – self-report measures of trait fear and anxiety are warranted. In closing, our review suggests several fruitful directions for future research. First, studies should strive to develop self-report measures of trait fear and trait anxiety based on the well-developed neurobiological and diagnostic research literatures. Such work should help to ascertain the extent to which the low to moderate correlations between trait fear and trait anxiety measures are due to artifacts, such as method covariance or content overlap, as opposed to genuine substantive overlap. Second, studies should investigate the neurobiological, cognitive, and behavioral correlates of these measures to better ascertain their etiological underpinnings. Functional brain imaging studies examining the patterns of neural activation in fear versus anxiety, at both state and trait levels, are especially needed. Third and finally, studies should investigate the relationships of both self-reported trait fear and trait anxiety with psychological disorders. Such work should help to ascertain which disorders currently housed within the broad anxiety disorders category are primarily disorders of trait fear as opposed to trait anxiety. Moreover, such work may bear implications for a number of conditions outside the current anxiety disorders category but in which abnormal (either high or low) levels of fear or anxiety play a prominent role, such as somatoform disorders, impulse control disorders, childhood disruptive disorders, and personality disorders. For example, although a number of authors (e.g., Lykken, 1995) have argued that psychopathy is characterized by low trait fear, this claim remains controversial, in part because the negative correlations between psychopathy and trait fear are often low in magnitude (Schmidt & Newman, 1999). Moreover, it is unclear which, if any, psychopathological syndromes are characterized by low trait anxiety (Watson & Clark, 1984). A clearer conceptual and empirical delineation between fear and anxiety will be an essential first step toward addressing these questions. Acknowledgements We thank Drs. Patricia A. Brennan, Michael Davis, Stephan Hamann, Stephen Nowicki, Jr., and Drew Westen for their feedback regarding this article. References1 Aggleton, J. P. (1992). The functional effects of amygdala lesions in humans: A comparison with findings from monkeys. In J. P. Aggleton (Ed.), The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction (pp. 485−503). New York: Wiley-Liss. Aguilar, R., Gil, L., Gray, J. A., Driscoll, P., Flint, J., Dawson, G. R., et al. (2003). Fearfulness and sex in F2 Roman rats: Male display more fear though both sexes share the same fearfulness traits. Physiology & Behavior, 78, 723−732. American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders (Revised), 4th ed. Washington, DC: Author. Antony, M. M., Orsillo, S. M., & Roemer, L. (Eds.). (2001). Practitioner's guide to empirically based measures of anxiety. New York: Kluwer Academic/Plenum. 1 * indicates references used in the meta-analysis. P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 Arrindell, W. (1980). Dimensional structure and psychopathology correlates of the fear survey schedule in a phobic population. Behavioral Research & Therapy, 18, 229−242. Barlow, D. H. (2002). Anxiety and its disorders: The nature and treatment of anxiety and panic. New York: Guilford Press. *Bates, H. D. (1971). Factorial structure and MMPI correlates of a fear survey schedule in a clinical population. Behaviour Research and Therapy, 9, 355−360. Beck, A. T., & Emery, G. (2005). Anxiety disorders and phobias: A cognitive perspective. New York: Basic Books. Berg, W. K., & Davis, M. (1985). Associative learning modifies startle reflexes at the lateral lemniscus. Behavioral Neuroscience, 99, 191−199. Bieling, P. J., Antony, M. M., & Swinson, R. P. (1998). The State-Trait Anxiety Inventory, Trait Version: Structure and content re-examined. Behavior Research and Therapy, 36, 777−788. Bishop, S., Duncan, J., Brett, M., & Lawrence, A. D. (2004). Prefrontal cortical function and anxiety: Controlling attention to threat-related stimuli. Nature Neuroscience, 7, 184−188. Blair, R. J. R. (2003). Neurobiological basis of psychopathy. British Journal of Psychiatry, 182, 5−7. Blanchard, D. C., & Blanchard, R. J. (1972). Innate and conditioned reactions to threat in rats with amygdaloid lesions. Journal of Comparative and Physiological Psychology, 81, 281−290. Blanchard, R. J., & Blanchard, D. C. (1990). An ethoexperimental analysis of defense, fear and anxiety. In N. McNaughton, & G. Andrews (Eds.), Anxiety (pp. 124−133). Dunedin: Otago University Press. *Blankstein, K. (1976). Relationship between Speilberger trait anxiety and Lykken social and physical trait anxiety. Journal of Clinical Psychology, 34, 781−782. Blundell, J., & Adamec, R. (2007). The NMDA receptor antagonist CPP blocks the effects of predator stress on pCREB in brain regions involved in fearful and anxious behavior. Brain Research, 1136, 59−76. Blundell, J., Adamec, R., & Burton, P. (2005). Role of NMDA receptors in the syndrome of behavioral changes produced by predator stress. Physiology & Behavior, 86, 233−243. Brown, J. S., Kalish, H. L., & Farber, I. E. (1951). Conditional fear as revealed by magnitude of startle response to an auditory stimulus. Journal of Experimental Psychology, 41, 317−328. Buss, A. H. (1962). Two anxiety factors in psychiatric patients. Journal of Abnormal and Social Psychology, 65, 426−427. *Byrne, B. (2000). Relationship between anxiety, fear, self esteem, and coping strategies in adolescence. Adolescence, 35(137), 201−215. *Carleton, R. N., Collimore, K. C., & Asmundson, G. J. G. (2007). Social anxiety and fear of negative evaluation: Construct validity of the BFNE-II. Journal of Anxiety Disorders, 21, 131−141. *Carsrud, A. L., & Carsrud, K. B. (1979). The relationship of sex role and levels of defensiveness to self-reports of fear and anxiety. Journal of Clinical Psychology, 35, 573−575. *Church, T. (1994). Relating the Tellegen and five-factor models of personality structure. Journal of Personality and Social Psychology, 67, 898−909. Cloninger, C. R., Przybeck, T. R., & Svrakic, D. M. (1991). Tridimensional Personality Questionnaire: US normative data. Psychological Reports, 69, 1047−1057. Cohen, H., Zohar, J., Matar, M. A., Zeev, K., Loewenthal, U., & Richter-Levin, G. (2004). Setting apart the affected: The use of behavioral criteria in animal models of post traumatic stress disorder. Neuropsychopharmacology, 29, 1962−1970. Cook, E. W., Davis, T. L., Hawk, L. W., Spence, E. L., & Gautier, C. H. (1992). Fearfulness and startle potentiation during aversive visual stimuli. Psychophysiology, 29, 633−645. Cooper, A. B., & Guynn, R. W. (2006). Transcription of fragments of lectures in 1948 by Harry Stack Sullivan. Psychiatry: Interpersonal and Biological Processes., 69, 110−112. Crawley, J. N. (1981). Neuropharmacologic specificity of a simple animal model for the behavioral actions of benzodiazipines. Pharmacology, Biochemistry, and Behavior, 15, 695−699. Crestani, F., Lorez, M., Baer, K., Essrich, C., Benke, D., Laurent, J. P., et al. (1999). Decreased GABAA-receptor clustering results in enhanced anxiety and a bias for threat cues. Nature Neuroscience, 2, 833−839. Davidson, R. J. (2002). Anxiety and affective style: Role of prefrontal cortex and amygdala. Biological Psychiatry, 51, 68−80. Davis, M. (1998). Are different parts of the extended amygdala involved in fear versus anxiety? Biological Psychiatry, 44, 1239−1247. Davis, M. (2006). Neural systems involved in fear and anxiety measured with fearpotentiated startle. Journal of Neuroscience, 22, 2343−2351. de Wit, S., Yutaka, K., Balleine, B. W., & Dickinson, A. (2006). Dorsomedial prefrontal cortex resolves response conflict in rats. Journal of Neuroscience, 26, 5224−5229. Degroot, A., & Nomikos, G. G. (2005). Fluoxetine disrupts the integration of anxiety and aversive memories. Neuropsychopharmacology, 30, 391−400. *Dien, J. (1999). Differential lateralization of trait anxiety and trait fearfulness: Evoked potential correlates. Personality and Individual Differences, 26, 333−356. Dupont, R. L., Rice, D. P., Miller, L. S., Shiraki, S. S., Rowland, C. R., & Harwood, H. J. (1996). Economic costs of anxiety disorders. Anxiety, 2, 167−172. Edinger, K. L., & Frye, C. A. (2006). Intrahippocampal administration of an androgen receptor agonist, flutamide, can increase anxiety-like behavior in intact and DHTreplaced male rats. Hormones and Behavior, 50, 216−222. Engels, A. S., Heller, W., Mohanty, A., Herrington, J. D., Banich, M. T., Webb, A. G., et al. (2007). Specificity of regional brain activity in anxiety types during emotion processing. Psychophysiology, 44, 352−363. Epstein, S. (1972). The nature of anxiety with emphasis upon its relationship to expectancy. In C. D. Spielberger (Ed.), Anxiety: Current trends in theory and research, Vol. 2. (pp. 291−337)New York: Academic Press. 135 Fendt, M., & Fanselow, M. S. (1999). The neuroanatomical and neurochemical basis of conditioned fear. Neuroscience and Biobehavioral Reviews, 23, 743−760. Fenz, W. D., & Epstein, S. (1965). Manifest anxiety: Unifactorial or multifactorial composition? Perceptual and Motor Skills, 20, 773−780. Funayama, E. S., Grillon, C., Davis, M., & Phelps, E. A. (2001). A double dissociation in the affective modulation of startle in humans: Effects of unilateral temporal lobectomy. Journal of Cognitive Neuroscience, 13, 721−729. Furmark, T., Fischer, H., Wik, G., Larsson, M., & Fredrikson, M. (1997). The amygdala and individual differences in human fear conditioning. Neuroreport, 8, 3957−3960. Geer, J. H. (1965). The development of a scale to measure fear. Behaviour Research and Therapy, 3, 45−53. Gewirtz, J. C., Falls, W. A., & Davis, M. (1997). Normal conditioned inhibition and extinction of freezing and fear-potentiated startle following electrolytic lesions of medical prefrontal cortex in rats. Behavioral Neuroscience, 111, 712−726. Gray, J. A. (1976). The behavioural inhibition system: A possible substrate for anxiety. In M. P. Feldman, & A. M. Broadhurst (Eds.), The theoretical and experimental bases of behaviour modification. London: Wiley. Gray, J. A. (1982). The neuropsychology of anxiety, 1st Ed. New York: Oxford University Press. Gray, J. A., & McNaughton, N. (2000). The neuropsychology of anxiety, 2nd Ed. New York: Oxford University Press. Grillon, C., & Ameli, R. (2001). Conditioned inhibition of fear-potentiated startle and skin conductance in humans. Psychophysiology, 38, 807−815. Grillon, C., Pellowski, M., Merikangas, K. R., & Davis, M. (1997). Darkness facilitates the acoustic startle reflex in humans. Biological Psychiatry, 42, 453−460. *Grossberg, J. M., & Wilson, H. K. (1965). A correlational comparison of the Wolpe-Lang Fear Survey Schedule and Taylor Manifest Anxiety Scale. Behaviour Research & Therapy, 3, 125−128. *Gullone, E., & King, N. J. (1992). Psychometric evaluation of a revised fear survey schedule for children and adolescents. Journal of Child Psychology and Psychiatry, 33, 987−998. Gur, R. C., Gur, R. E., Skolnick, B. E., Resnick, S. M., Silver, F. L., Chowluk, J., et al. (1988). Effects of task difficulty on regional cerebral blood flow: Relationships with anxiety and performance. Psychophysiology, 25, 392−399. *Hagopian, L. P., & Ollendick, T. H. (1996). Behavioral inhibition and anxiety sensitivity: A reanalysis. Personality and Individual Differences, 21, 247−252. Hamm, A. O., & Vaitl, D. (1996). Affective learning: Awareness and aversion. Psychophysiology, 33, 698−710. Hayden, E. P., Dougherty, L. R., Maloney, B., Durbin, E. C., Olino, T. M., Nurnberger, J. I., et al. (2007). Temperamental fearfulness in childhood and the serotonin transporter promoter region polymorphism: A multimethod association study. Psychiatric Genetics, 17, 135−142. Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando: Academic Press. *Helfritz, L. E., & Stanford, M. S. (2006). Personality and psychopathology in an impulsive aggressive college sample. Aggressive Behavior, 32, 28−37. Heller, W., Nitschke, J. B., Etienne, M. A., & Miller, G. A. (1997). Patterns of regional brain activity differentiate types of anxiety. Journal of Abnormal Psychology, 106, 376−385. *Hersen, M. (1971). Fear scale norms for an in-patient population. Journal of Clinical Psychology, 27, 375−378. Hitchcock, J. M., & Davis, M. (1991). The efferent pathway of the amygdala involved in conditioned stimulus: Effect of lesions of the amygdala. Behavioral Neuroscience, 105, 826−842. *Hodges, W. F., & Felling, J. P. (1970). Types of stressful situations and their relation to trait anxiety and sex. Journal of Consulting and Clinical Psychology, 34, 333−337. Izard, C. E., & Ackerman, B. P. (2000). Motivational, organizational, and regulatory functions of discrete emotions. In M. Lewis, & J. M. Haviland-Jones (Eds.), Handbook of emotions (pp. 253−264)., 2nd Ed. New York: Guilford Press. Jovanovic, T., Norrholm, S. D., Keyes, M., Fiallos, A., Jovanovic, S., Myers, K. M., et al. (2006). Contingency awareness and fear inhibition in a human fear-potentiated startle paradigm. Behavioral Neuroscience, 120, 995−1004. Kalin, N. H., Shelton, S. E., & Davidson, R. J. (2004). The role of the central nucleus of the amygdala in mediating fear and anxiety in the primate. Journal of Neuroscience, 24, 5506−5515. Kalisch, R., Salome, N., Platzer, S., Wigger, A., Czisch, M., Sommer, W., et al. (2004). High trait anxiety and hyporeactivity to stress of the dorsomedial prefrontal cortex: A combined phMRI and Fos study in rats. Neuroimage, 23, 382−391. Kendler, K. S., Prescott, C. A., Myers, J., & Neale, M. C. (2003). The structure of genetic and environmental risk factors for common psychiatric and substance use disorders in men and women. Archives of General Psychiatry, 60, 929−937. *Kent, G., & Keohane, S. (2001). Social anxiety and disfigurement: The moderating effects of fear of negative evaluation and past experience. British Journal of Clinical Psychology, 40, 23−34. Kessler, R. C., Chiu, W. T., Demler, O., & Walters, E. E. (2005). Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the national comorbidity survey replication. Archives of General Psychiatry, 62, 617−627. *Kilpatrick, D. G., & McLeod, P. G. (1973). Trait anxiety and fearfulness. Social Behavior and Personality: An International Journal, 2, 119−122. *Kilpatrick, D. G., Sutker, P. B., Roitzsch, J. C., & Mason, R. L. (1975). Self-reported fears and electrodermal responsiveness of high and low trait anxious subjects to fear of failure and other stressors. Social Behavior and Personality, 3, 205−211. King, N. J., & Campbell, I. M. (1986). A manifest anxiety scale from MMPI-168. Journal of Clinical Psychology, 42, 748−751. *King, N. J., Gullone, E., & Ollendick, T. H. (1992). Manifest anxiety and fearfulness in children and adolescents. Journal of Genetic Psychology, 153, 63−74. Klorman, R. (1974). Habituation of fear: Effects of intensity and stimulus order. Psychophysiology, 11, 15−26. 136 P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 *Kocovski, N. L., & Endler, N. S. (2000). Social anxiety, self-regulation, and fear of negative evaluation. European Journal of Personality, 14, 347−358. *Kogan, J. N., & Edelstein, B. A. (2004). Modification and psychometric examination of a self-report measure of fear in older adults. Journal of Anxiety Disorders, 18, 397−408. Krueger, R. F. (1999). The structure of common mental disorders. Archives of General Psychiatry, 56, 921−926. Lang, P. J., Davis, M., & Öhman, A. (2000). Fear and anxiety: Animal models and human cognitive psychophysiology. Journal of Affective Disorders, 61, 137−159. *Lang, P. J., & Lazovik, A. D. (1963). Experimental desensitization of a phobia. Journal of Abnormal and Social Psychology, 66, 519−525. LeDoux, J. E., Cicchetti, P., Xagoraris, A., & Romanski, L. M. (1990). The lateral amygdaloid nucleus: Sensory interface of the amygdala in fear conditioning. Journal of Neuroscience, 10, 1062−1069. LeDoux, J. E., Iwata, J., Cicchetti, P., & Reis, D. J. (1988). Different projections of the central amygdaloid nucleus mediate autonomic and behavioral correlates of conditioned fear. Journal of Neuroscience, 8, 2517−2529. *Lerner, J. S., & Keltner, D. (2001). Fear, anger, and risk. Journal of Personality and Social Psychology, 81, 146−159. Lilienfeld, S. O. (1990). Development and preliminary validation of a self-report measure of psychopathic personality. Doctoral dissertation: University of Minnesota, Minneapolis. Lilienfeld, S. O., & Andrews, B. P. (1996). Development and preliminary validation of a self report measure of psychopathic personality traits in noncriminal populations. Journal of Personality Assessment, 66, 488−524. Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Thousand Oaks, CA: Sage Publications. Lykken, D. T. (1995). The antisocial personalities. New York: Lawrence Erlbaum. *Lykken, D. T., Tellegen, A., & Katzenmeyer, C. G. (1973). Manual for the Activity Preference Questionnaire. Report No. PR-73-4, Reports from the Research Laboratories. Minneapolis: Department of Psychiatry, University of Minnesota. Macleod, C., & Rutherford, E. M. (1992). Anxiety and the selective processing of emotional information: Mediating roles of awareness, trait and state variables, and personal relevance of stimulus materials. Behavior Research and Therapy, 30, 479−491. *Manosevitz, M., & Lanyon, R. I. (1965). Fear survey schedule: A normative study. Psychological Reports, 17, 699−703. Marks, I. M., & Matthews, A. M. (1979). Brief standard self-rating for phobic patients. Behavior Research & Therapy, 17, 263−267. Martinot, J. H., Allilaire, J. F., Mazolyer, B. M., Hantouche, E., Huret, J. D., Legaut-Demare, F., et al. (1990). Obsessive–compulsive disorder: A clinical neuropsychological and positron emission tomography study. Acta Psychiatrica Scandanavica, 82, 233−242. Matt, G. E., & Cook, T. D. (1994). Threats to the validity of research synthesis. In H. Cooper, & L. V. Hedges (Eds.), The handbook of research synthesis (pp. 503−520). New York: Russell Sage Foundation. *Mattick, R. P., & Clarke, J. C. (1998). Development and validation of measures of social phobia, scrutiny fear, and social interaction anxiety. Behaviour Research and Therapy, 36, 455−470. McNaughton, N., & Corr, P. J. (2004). A two-dimensional neuropsychology of defense: Fear/anxiety and defensive distance. Neuroscience & Biobehavioral Review, 28, 285−305. McNaughton, N., & Gray, J. A. (2000). Anxiolytic action on the behavioural inhibition system implies multiple types of arousal contribute to anxiety. Journal of Anxiety Disorders, 61, 161−176. McNaughton, N., Kocsis, B., & Hajos, M. (2007). Elicited hippocampal theta rhythm: A screen for anxiolytic and procognitive drugs through changes in hippocampal function? Behavioural Pharmacology, 18, 329−346. *Mellon, R. (2000). A Greek-language inventory of fears: Psychometric properties and factor structure of self-reports of fears on the Hellenic Fear Survey Schedule. Journal of Psychopathology and Behavioral Assessment, 22, 123−140. Messick, S. (1995). Validity of psychological assessment: Validation of inferences from persons' responses and performances as scientific inquiry into score meaning. American Psychologist, 50, 741−749. Milad, M. R., Quinn, B. T., Pitman, R. K., & Orr, S. P. (2005). Thickness of ventromedial prefrontal cortex in humans is correlated with extinction memory. Proceedings of the National Academy of Sciences, 102, 10706−10711. Milad, M. R., Wright, C., Orr, S., Pitman, R., Quirk, G., & Rauch, S. (2007). Recall of fear extinction in humans activates the ventromedial prefrontal cortex and hippocampus in concert. Biological Psychiatry, 62, 446−454. Miller, K. A., Garner, J. P., & Mench, J. A. (2006). Is fearfulness a trait that can be measured with behavioural tests? A validation of four fear tests for Japanese quail. Animal Behaviour, 71, 1323−1334. Mineka, S. (1979). The role of fear in theories of avoidance learning, flooding, and extinction. Psychological Bulletin, 86, 985−1010. Morgan, M. A., Romanski, L. M., & LeDoux, J. E. (1993). Extinction of emotional learning: Contribution of medial prefrontal cortex. Neuroscience Letters, 163, 109−113. Morinaga, K., Akiyoshi, J., Matsushita, H., Ichioka, S., Tanaka, Y., Tsuru, J., et al. (2007). Biological Psychology, 74, 34−38. Morrissette, S. B., Tull, M. T., Gulliver, S. B., Kamholz, B. W., & Zimering, R. T. (2007). Anxiety, anxiety disorders, tobacco use, and nicotine: A critical review of interrelationships. Psychological Bulletin, 133, 245−272. *Muris, P., Merckelbach, H., Mayer, B., van Brakel, A., Thissen, S., Moulaert, V., et al. (1998). The Screen for Child Anxiety Emotional Disorders (SCARED) and traditional childhood anxiety measures. Journal of Behavior Therapy and Experimental Psychiatry, 29, 327−339. *Muris, P., Meesters, C., Mayer, B., Bogie, N., Luijten, M., Geebelen, E., et al. (2003). The Koala Fear Questionnaire: A standardized self-report scale for assessing fears and fearfulness in pre-school and primary school children. Behaviour Research and Therapy, 5, 597−617. Nader, K., & LeDoux, J. (1999). The dopaminergic modulation of fear: Quinipirole impairs the recall of emotional memories in rats. Behavioral Neuroscience, 113, 152−165. *Newman, J. P., & Schmidt, W. A. (1998). Passive avoidance in psychopathic offenders: A replication and extension. Journal of Abnormal Psychology, 107, 527−532. Oei, T. P. S., Evans, L., & Crook, G. (1990). Utility and validity of the STAI with anxiety disorder patients. British Journal of Clinical Psychology, 29, 429−432. Oei, T. P. S., Moylan, A., & Evans, L. (1991). Validity and clinical utility of the Fear Questionnaire for anxiety patients. Psychological Assessment, 3, 391−397. Öhman, A. (1986). Face the beast and fear the face: Animal and social fears as prototypes for evolutionary analyses of emotion. Psychophysiology, 23, 123−145. Öhman, A. (1993). Fear and anxiety as emotional phenomena: Clinical phenomenology, evolutionary perspectives, and information-processing mechanisms. In M. Lewis, & J. M. Haviland (Eds.), Handbook of emotions (pp. 511−536). New York, NY: Guilford Press. Öhman, A. (2008). Fear and anxiety: Overlaps and dissociations. In M. Lewis, J. M. Haviland-Jones, & L. F. Barrett (Eds.), Handbook of emotions (pp. 709−729)., 3rd Ed. New York, NY: Guilford Press. *Ollendick, T. H. (1983). Reliability and Validity of the Revised Fear Survey Schedule for Children (FSSC-R). Behavior, Research, and Therapy, 21, 684−692. Onaivi, E. S., & Martin, B. R. (1989). Neuropharmacological and physiological validation of a computer-controlled two-compartment black and white box for the assessment of anxiety. Progress in Neuropsychopharmacology and Biological Psychiatry, 13, 963−976. *Peleg-Popko, O., & Dar, R. (2001). Marital quality, family patterns, and children's fears and social anxiety. Contemporary Family Therapy, 23, 465−487. *Perkins, A. M., Kemp, S. E., & Corr, P. J. (2007). Fear and anxiety as separable emotions: An investigation of the revised reinforcement sensitivity theory of personality. Emotion, 7, 252−261. Pine, D. S., Fyer, A., Grun, J., Phelps, E. A., Szeszko, P. R., Koda, V., et al. (2001). Methods for developmental studies of fear conditioning circuitry. Biological Psychiatry, 50, 225−228. Porges, S. W. (1995). Orienting in a defensive world: Mammalian modifications to our evolutionary heritage. A polyvagal theory. Psychophysiology, 32, 301−318. Poythress, N. G., Edens, J. F., Landfield, K., Lilienfeld, S. O., Skeem, J. L., & Douglas, K. S. (2008). A critique of Carver and White's (1994) Behavioral Inhibition Scale (BIS) for investigating Lykken's theory of primary psychopathy. Personality and Individual Differences, 45, 269−275. Quirk, G. J., Russo, G. K., Barron, J. L., & LeBron, K. (2000). The role of the ventro-medial prefrontal cortex in the recovery of extinguished fear. Journal of Neuroscience, 20, 6225−6231. Reiman, E. M., Raichle, M. E., Robins, E., Mintum, M. A., Fusselman, M. J., Fox, P. T., et al. (1989). Neuroanatomical correlates of lactate-induced anxiety attack. Archives of General Psychiatry, 46, 493−500. Rhudy, J. L., & Meagher, M. W. (2000). Fear and anxiety: Divergent effects on human pain thresholds. Pain, 84, 65−75. Rosenthal, R. (1979). Comment: Assumptions and procedures in the file drawer problem. Psychological Bulletin, 86, 638−641. Ruiz-Padial, E., & Vila, J. (2007). Fearful and sexual pictures not consciously seen modulate the startle reflex in human beings. Biological Psychiatry, 61, 996−1001. *Scherer, M. W., & Nakamura, C. Y. (1968). A Fear Survey Schedule for Children (FSSC): A factor analytic comparison with manifest anxiety. Behaviour, Research, and Therapy, 6, 173−182. *Schmidt, N. B., & Mallott, M. (2005). Evaluating anxiety sensitivity and other fundamental sensitivities predicting anxiety symptoms and fearful responding to a biological challenge. Behavior Research and Therapy, 44, 1681−1688. *Schmidt, W. A., & Newman, J. P. (1999). Are all psychopathic individuals low anxious? Journal of Abnormal Psychology, 108, 353−358. *Schroeder, H., & Craine, L. (1971). Relationships among measures of fear and anxiety for snake phobics. Journal of Consulting and Clinical Psychology, 36, 443. *Sellbom, M., & Ben-Porath, Y. S. (2005). Mapping the MMPI-2 restructured clinical scales onto normal personality traits: Evidence of construct validity. Journal of Personality Assessment, 85, 179−187. Shadish, W. R., Cook, T. D., & Campbell, D. T. (2001). Experimental and quasiexperimental design for generalized causal inference. Boston: Houghton Mifflin. *Shore, G. N., & Rapport, M. D. (1998). The Fear Survey Schedule for Children—Revised (FSSC-HI): Ethnocultural variations in children's fearfulness. Journal of Anxiety Disorders, 12, 437−461. *Silverman, W. K., Fliesig, W., Rabian, B., & Peterson, R. (1991). Childhood anxiety sensitivity index. Journal of Clinical Child Psychology, 20, 162−168. *Stanley, M. A., Beck, J. G., & Zebb, B. J. (1996). Psychometric properties of four anxiety measures in older adults. Behaviour Research and Therapy, 34, 827−838. Sierra-Mercado, D., Corcoran, K. A., Lebrón-Milad, K., & Quirk, G. J. (2006). Inactivation of the ventromedial prefrontal cortex reduces expression of conditioned fear and impairs subsequent recall of extinction. European Journal of Neuroscience, 24, 1751−1758. Sotres-Bayon, F., Cain, C. K., & LeDoux, J. E. (2006). Brain mechanisms of fear extinction: Historical perspectives on the contribution of prefrontal cortex. Biological Psychiatry, 60, 329−336. Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., & Jacobs, G. A. (1983). Manual for the State-Trait Anxiety Inventory (Form Y). Palo Alto: Mind Garden. Straub, T., Schmidt, S., Weiss, T., Mentzel, H., & Miltner, W. H. R. (2009). Dynamic activation of the anterior cingulated cortex during anticipatory anxiety. Neuroimage, 44, 975−981. P. Sylvers et al. / Clinical Psychology Review 31 (2011) 122–137 Stravynski, A., Basoglu, M., Marks, M., Sengun, S., & Marks, I. M. (1995). The distinctiveness of phobias: A discriminant analysis of fears. Journal of Anxiety Disorders, 9, 89−101. Tang, J., & Gibson, S. J. (2005). A psychophysical evaluation of the relationship between trait anxiety, pain perception, and induced state anxiety. Journal of Pain, 6, 612−619. Taylor, J. A. (1953). A personality scale of manifest anxiety. Journal of Abnormal and Social Psychology, 48, 285−290. Tellegen, A. (1982). Brief manual for the Multidimensional Personality Questionnaire.: University of Minnesota. Tellegen, A. (1985). Structures of mood and personality and their relevance to assessing anxiety, with an emphasis on self-report. In T. A. Hussain, & J. D. Maser (Eds.), Anxiety and the anxiety disorders (pp. 681−706). Hillsdale, NJ: Lawrence Erlbaum. Tellegen, A., & Waller, N. G. (2007). Exploring personality through test construction: Development of the Multidimensional Personality Questionnaire. Minneapolis: University of Minnesota Press. Trull, T. J., & Hillerbrand, E. (1990). Psychometric properties and factor structure of the Fear Questionnaire phobia subscale items in two normative samples. Journal of Psychopathology and Behavioral Assessment, 12, 285−297. *Uzieblo, K., Verscuerea, B., & Crombez, G. (2007). The Psychopathic Personality Inventory: Construct validity of the two-factor structure. Personality and Individual Differences, 43, 657−667. Vendruscolo, L. F., Takahashi, R. N., Brüske, G. R., & Ramos, A. (2003). Evaluation of the anxiolytic-like effect of NKP608, a NK1-receptor antagonist, in two rat strains that differ in anxiety-related behaviors. Psychopharmacology, 170, 287−293. Vrana, S. R., & Lang, P. J. (1990). Fear imagery and the startle-probe reflex. Journal of Abnormal Psychology, 99, 189−197. Walker, D. L., & Davis, M. (1997). Anxiogenic effects of high illumination levels assessed with the acoustic startle paradigm. Biological Psychiatry, 42, 461−471. *Waller, N. G., Lilienfeld, S. O., Tellegen, A., & Lykken, D. (1991). The Tridimensional Personality Questionnaire: Structural validity and comparison with the Multidimensional Personality Questionnaire. Multivariate Behavioral Research, 26, 1−23. *Walters, G. D. (2001). State-trait anxiety and existential fear: An empirical analysis. Personality and Individual Differences, 30, 1345−1352. Watson, D. (2005). Rethinking the mood and anxiety disorders: A quantitative hierarchical model for DSM-V. Journal of Abnormal Psychology, 114, 522−536. 137 Watson, D., & Clark, L. A. (1984). Negative affectivity: The disposition to experience aversive emotional states. Psychological Bulletin, 96, 465−490. Watson, D., Clark, L. A., & Harkness, A. R. (1994). Structures of personality and their relevance to psychopathology. Journal of Abnormal Psychology, 103, 18−31. *Weeks, J. W., Heimberg, R. G., Fresco, D. M., Hart, T. A., Turk, C. L., Schneier, F. R., et al. (2005). Empirical validation and psychometric evaluation of the brief fear of negative evaluation scale in patients with social anxiety disorder. Psychological Assessment, 17, 179−190. Whalen, P. J., Rauch, S. L., Etcoff, N. L., McInerney, S. C., Lee, M. B., & Jenike, M. A. (1998). Masked presentations of emotional facial expressions modulate amygdala activity without explicit knowledge. Journal of Neuroscience, 18, 411−418. White, T. L., & Depue, R. A. (1999). Differential association of traits of fear and anxiety with norepinephrine- and dark-induced pupil reactivity. Journal of Personality and Social Psychology, 77, 863−877. Wilensky, A. E., Schafe, G. E., Kristensen, M. P., & LeDoux, J. E. (2006). Rethinking the fear circuit: The central nucleus of the amygdala is required for the acquisition, consolidation, and expression of Pavlovian fear conditioning. Journal of Neuroscience, 26, 12387−12396. Williamson, D. E., Coleman, K., Bacanu, S., Delvin, B., Rogers, J., Ryan, N. D., et al. (2003). Heritability of fearful–anxious endophenotypes in infant rhesus macaques: A preliminary report. Biological Psychiatry, 53, 284−291. *Wilson, K. A., & Hayward, C. (2006). Unique contributions of anxiety sensitivity to avoidance: A prospective study in adolescents. Behaviour Research and Therapy, 44, 601−609. Winslow, J. T., Noble, P. L., & Davis, M. (2007). Modulation of fear-potentiated startle and vocalizations in juvenile rhesus monkeys by morphine, diazepam, and buspirone. Biological Psychiatry, 61, 389−395. Wolpe, J. (1987). The promotion of scientific psychotherapy: A long voyage. In J. K. Zeig (Ed.), The evolution of psychotherapy (pp. 133−142). New York: Routledge. Wolpe, J., & Lang, P. J. (1964). A fear survey schedule for use in behaviour therapy. Behavior Research and Therapy, 2, 27−30. Zinbarg, R. E., & Mineka, S. (2007). Is emotion regulation a useful construct that adds to the explanatory power of learning models of anxiety disorders or a new label for old constructs. American Psychologist, 62, 259.